Telematics Wireless Architecture and Application (2010) GOOD Lecture

Telematics Communication
Technologies and
Vehicular Networks:
Wireless Architectures
and Applications
Chung-Ming Huang
National Cheng Kung University, Tainan, Taiwan, R.O.C.
Yuh-Shyan Chen
National Taipei University, Taipei, Taiwan, R.O.C.
InformatIon scIence reference

Hershey • New York
Director of Editorial Content: Kristin Klinger
Senior Managing Editor: Jamie Snavely
Assistant Managing Editor: Michael Brehm
Publishing Assistant: Sean Woznicki
Typesetter: Michael Brehm, Carole Coulson
Cover Design: Lisa Tosheff
Printed at: Yurchak Printing Inc.
Published in the United States of America by

Information Science Reference (an imprint of IGI Global)
701 E. Chocolate Avenue
Hershey PA 17033
Tel: 717-533-8845
Fax: 717-533-8661
E-mail: cust@igi-global.com
Web site: http://www.igi-global.com/reference
Copyright © 2010 by IGI Global. All rights reserved. No part of this publication may be reproduced, stored or distributed in
any form or by any means, electronic or mechanical, including photocopying, without written permission from the publisher.
Product or company names used in this set are for identification purposes only. Inclusion of the names of the products or
companies does not indicate a claim of ownership by IGI Global of the trademark or registered trademark.
Library of Congress Cataloging-in-Publication Data
Telematics communication technologies and vehicular networks : wireless architectures and applications / Chung-Ming Huang
and Yuh-Shyan Chen, editors.
p. cm.
Summary: "This book examines critical issues involved with telematics such as vehicular network infrastructure, vehicular
network communication protocols, and vehicular services and applications"--Provided by publisher.
Includes bibliographical references and index.
ISBN 978-1-60566-840-6 (hardcover) -- ISBN 978-1-60566-841-3 (ebook) 1. Automotive telematics. I. Huang, Chung-
Ming, 1961- II. Chen, Yuh-Shyan.
TL272.58.T45 2010
629.2'77--dc22
2009034903
British Cataloguing in Publication Data

A Cataloguing in Publication record for this book is available from the British Library.
All work contributed to this book is new, previously-unpublished material. The views expressed in this book are those of the
authors, but not necessarily of the publisher.
List of Reviewers
Chih-Yung Chang, Tamkang University, Taipei, Taiwan, R.O.C
Ruay-Shiung Chang, National Dong Hwa University, Hualien, Taiwan, R.O.C.
Tzung-Shi Chen, National University of Tainan, Tainan, Taiwan, R.O.C.
Wei-Kuo Chiang, National Chung Cheng University, Chia-Yi, Taiwan, R.O.C.
Li-Der Chou, National Central University, Chungli, Toayuan, Taiwan, R.O.C.
Chyi-Ren Dow, Feng Chia University, Taichung, Taiwan, R.O.C.
Ren-Hung Hwang, National Chung Cheng University, Chia-Yi, Taiwan, R.O.C.
Guan-Ling Lee, National Dong Hwa University, Hualien, Taiwan, R.O.C.
Wan-Jiun Liao, National Taiwan University, Taipei, Taiwan, R.O.C.
Tsung-Nan Lin, National Taiwan University, Taipei, Taiwan, R.O.C.
Jen-Yi Pan, National Chung Cheng University, Chia-Yi, Taiwan, R.O.C.
Shiao-Li Tsao, National Chiao Tung University, Hsinchu, Taiwan, R.O.C.
Table of Contents
Foreword ............................................................................................................................................ xix
Preface ................................................................................................................................................ xxi
Acknowledgment ............................................................................................................................... xxv
Section 1
Introduction of Vehicular Networks and Intelligent Transporation Systems
Chapter 1
Introduction of Vehicular Network Architectures ................................................................................... 1
Ming-Chiao Chen, National Taitung University, Taitung, Taiwan, R.O.C.
Teng-Wen Chang, National Taiwan University, Taipei, Taiwan, R.O.C.
Chapter 2
Introduction of Vehicular Network Applications .................................................................................. 15
Yao-Chung Chang, National Taitung University, Taiwan, R.O.C.
Chapter 3
Introduction to ITS and NTCIP ............................................................................................................ 32
Da-Jie Lin, Feng Chia University, Taiwan, R.O.C.
Chyi-Ren Dow, Feng Chia University, Taiwan, R.O.C.
Section 2
Embedded System Architecture and Communication Protocols
Chapter 4
Vehicular Embedded System Architecture............................................................................................ 58
Chung-Ping Young, National Cheng Kung University, Taiwan, R.O.C.
Chapter 5
Data Communications Inside Vehicular Environments ........................................................................ 74
Cheng-Min Lin, Nan Kai University of Technology, Taiwan, R.O.C.
Tzong-Jye Liu, Feng Chia University, Taiwan, R.O.C.
Chapter 6
Wireless Access in Vehicular Environments ......................................................................................... 90
Ching-Wen Chen, Feng Chia University, Taiwan, R.O.C.
Section 3
Location Based Services
Chapter 7
Introduction To Global Satellite Positioning System (GPS) ............................................................... 108
Jenq-Muh Hsu, National Chiayi University, Chiayi, Taiwan, R.O.C.
Chapter 8
Vehicle Location and Navigation Systems ......................................................................................... 119
Ben-Jye Chang, National Yunlin University of Science and Technology, Yunlin, Taiwan, R.O.C.
Chapter 9
Design and Implementation of Vehicle Navigation Systems .............................................................. 131
Min-Xiou Chen, National Dong-Hwa University, Hualien, Taiwan, R.O.C.
Section 4
Integrated Vehicular Application
Chapter 10
Vehicular Metropolitan Area Network Systems Architecture: The WiMAX Network
Reference Model ................................................................................................................................. 144
Cheng Hsuan Cho, National Chung Cheng University, Taiwan, R.O.C.
Jen-Yi Pan, National Chung Cheng University, Taiwan, R.O.C.
Chapter 11
Interworking of IP Multimedia Subsystem and Vehicular Communication Gateway ........................ 160
Wei-Kuo Chiang, National Chung Cheng University, Chiaya, Taiwan, R.O.C.
An-Nie Ren, National Chung Cheng University, Chiaya, Taiwan, R.O.C.
Section 5
Vehicular Ad Hoc Networks and Delay Tolerant Vehicular Networks
Chapter 12
MAC Protocols in Vehicular Ad Hoc Networks ................................................................................. 183
Chih-Yung Chang, Tamkang University, Taiwan, R.O.C.
Chapter 13
Routing Protocol in Vehicular Ad Hoc Networks ............................................................................... 206
Yuh-Shyan Chen, National Taipei University, Taipei, Taiwan, R.O.C.
Yun-Wei Lin, National Taipei University, Taipei, Taiwan, R.O.C.
Chapter 14
Applications in Vehicular Ad Hoc Networks ...................................................................................... 229
Hua-Wen Tsai, National University of Tainan, Tainan, Taiwan, R.O.C.
Yi-Shiang Chang, National University of Tainan, Tainan, Taiwan, R.O.C.
Chapter 15
DTN Technologies for Vehicular Networks........................................................................................ 252
Kun-Chan Lan, National Cheng Kung University, Tainan, Taiwan, R.O.C.
Section 6
Management and Traffic Control
Chapter 16
Simple Transporation Management Framework................................................................................. 271
Chapter 17
Vehicular System Management Architecture and Application Platform ............................................ 290
Teng-Wen Chang, National Taiwan University of Science and Technology, Taiwan, R.O.C.
Jiann-Liang Chen, National Taiwan University of Science and Technology, Taiwan, R.O.C.
Chapter 18
Remote Vehicular System Management Functions and Information Structure .................................. 310
Chapter 19
Using Wireless Mesh Network for Traffic Control ............................................................................. 331
Section 7
Mobility Model, Simulation, and Security
Chapter 20
Mobility Models of Vehicular Networks ............................................................................................ 348
Chapter 21
MOVE: A Practical Simulator for Mobility Model in VANET .......................................................... 355
Chapter 22
Security Attacks of Vehicular Networks ............................................................................................. 369
Jen-Chun Chang, National Taipei University, Taiwan, R.O.C.
Chun-I Fan, National Sun Yat-sen University, Taiwan, R.O.C.
Ruei-Hau Hsu, National Sun Yat-sen University, Taiwan, R.O.C.
Compilation of References ............................................................................................................... 380
About the Contributors .................................................................................................................... 398
Index ................................................................................................................................................... 405

Detailed Table of Contents
Foreword ............................................................................................................................................ xix
Preface ................................................................................................................................................ xxi
Acknowledgment ............................................................................................................................... xxv
Section 1
Introduction of Vehicular Networks and Intelligent Transporation Systems
Chapter 1
Introduction of Vehicular Network Architectures ................................................................................... 1
Ming-Chiao Chen, National Taitung University, Taitung, Taiwan, R.O.C.
Teng-Wen Chang, National Taiwan University, Taipei, Taiwan, R.O.C.
A vehicular network organizes and connects vehicles with each other, and with mobile and fixed-locations
resources. This chapter discusses the architectures in the vehicular network environment. We intro-
duce the overview of in-vehicle and out-vehicle network architectures. An automobile in an in-vehicle
network adopts four vehicle bus protocols, CAN (Controller Area Network), LIN (Local Interconnect
Network), MOST (Media Oriented Systems Transport) and FlexRay. However, these protocols cannot
intercommunicate with each other. Therefore, the OSEK operating system is designed as standard soft-
ware architecture for the various ECUs (Electronic Control Units). In the out-vehicle network, the OBU
(On Board Unit) in the automobile can communicate with the infrastructure via the Internet. We discuss
next-generation vehicular network architecture, the modern in-vehicle networks, on-board computers
and the Internet, mobile telecommunications and telematics applications in the ground vehicles, and
finally, we introduce future desired features. This chapter discusses the architectures in vehicular network
environment. Section 1.1 introduces the overview of in-vehicle and out-vehicle network architectures.
Section 1.2 describes in-vehicle network architecture for disaster communication network by combining
various automotive bus protocols. Section 1.3 describes the out-vehicle network architecture for disaster
communication network by combining various wireless LANs. Section 1.4 discusses next-generation
vehicular network architecture, the modern in-vehicle networks, on-board computers and the Internet,
mobile telecommunications and telematics applications in the ground vehicles, and introduces future
desired features.
Chapter 2
Introduction of Vehicular Network Applications .................................................................................. 15
Yao-Chung Chang, National Taitung University, Taiwan, R.O.C.
Information and Communication Technology (ICT) is concerned with all the technologies that manage,
process, and communicate information. It is also named as telematics, combining two words: telecom-
munications and informatics, which is widely used in the application of Global Positioning System
technology integrated with computers and in the mobile communications technology for automotive
navigation systems. Table 2.1 and Table 2.2 respectively list the telemetric applications from user’s
point of view and the practical applications of vehicular telematics. Four applications of the vehicular
network are discussed in this chapter. Section 2.1 introduces the vehicular network application services.
Section 2.2 discusses the vehicular network application management. Section 2.3 provides the platform
technologies of vehicular network application. Finally, future vehicular network application and deploy-
ments are presented in the Section 2.4.
Chapter 3
Introduction to ITS and NTCIP ............................................................................................................ 32
Da-Jie Lin, Feng Chia University, Taiwan, R.O.C.
Intelligent Transportation Systems (ITS) combines high technology and improvements in information
systems, communication, sensors, and relevant mathematical methods with the conventional world of
surface transportation infrastructure to increase the capacity of transportation systems and to improve
the level of services. There are four major goals of ITS, including safety, environmental protection, effi-
ciency, and economy. NTCIP (NTCIP Standard 9001, 2002; DISA et al., 1997) is a set of communications
protocols and data definition standards designed for various needs of ITS services and applications. The
key goals of the NTCIP open-standards effort are interoperability and interchangeability. Interoperability
refers to the ability for multiple devices to work together as a single system and interchangeability refers
to the ability to use multiple brands of a device on the same communications channel. Accompanying the
social and economic development, traffic congestion and delay have become major issues in most areas
around the world. How to use readily available technologies to increase the capacity of transportation
systems and to improve the level of service has become one of major solutions to solve transportation
problems that people are facing. This is the motivation of Intelligent Transportation Systems develop-
ment. NTCIP is a set of communications protocols and data definition standards designed for various
needs of ITS services and applications. These standards are intended to handle these needs in the two
areas: Center-to-Field (C2F) and Center-to-Center (C2C) communications.
Section 2
Embedded System Architecture and Communication Protocols
Chapter 4
Vehicular Embedded System Architecture............................................................................................ 58
Chung-Ping Young, National Cheng Kung University, Taiwan, R.O.C.
Transportation of humans and objects have been playing an important role in our daily lives since civiliza-
tion first formed and needed new means of reaching destinations. The invention of efficient transportation
greatly reduced the time and labor once required and in addition largely extended the living environment
that people can reach. The more time and labor for transportation is saved, the more leisure time people
will have. Animal-power or natural resources have been the driving force of transportation for a long
time. After the steam engine was invented, the automobile started a new era. The mass production of the
Ford model T created the modern automobile industry and made the automobile more affordable. The
basic structure of the automobile has not changed much, but evolving technologies has kept improving
its functions and performance. The construction of traffic networks and mass production of automobiles
have made the automobile the most important land based transportation carrier. The usage of automobiles
is usually associated with the growth of economy and industry of a nation, so the population ratio that
owns automobiles in a developed country is larger than that in a developing country. When the economy
grows, vehicle as a transportation tool becomes more affordable and popular, for instance China or In-
dia. When people use automobiles in their daily lives, they demand not only mobility, but also safety,
comfort and convenience. These are some design factors that manufacturers have to put into aspect when
enhancing functions by introducing and developing new technologies.
Chapter 5
Data Communications Inside Vehicular Environments ........................................................................ 74
Cheng-Min Lin, Nan Kai University of Technology, Taiwan, R.O.C.
ZigBee is based on IEEE 802.15.4 which specifies the physical layer and medium access control (MAC)
for low-cost and low-power LR-WPAN. The technology can be applied in intelligent key, A/C opera-
tion and steering wheel inside vehicles. There are two types of devices in ZigBee, FFD and RFD. A
FFD can communicate with RFDs and other FFDs, while a RFD can only communicate with a FFD.
In ZigBee physical layer, it follows IEEE 802.15.4 standard and operates in unlicensed RF worldwide
(2.4GHz global, 915MHz Americas or 868 MHz Europe). A superframe contained an active portion and
an inactive portion is used in the MAC layer of ZigBee. The active portion includes CAP and CFP. In
the inactive partition, the coordinator can enter sleep mode to save its power. Three main topologies of
ZigBee are star, mesh, and tree. However, ZigBee is successfully produced into a low-cost controller
applied for automotive applications, including vehicle control and status monitoring. According to the
forecast of ON World in 2005 (ON WORLD, 2009) , the deployed wireless sensing network nodes will
increase to 127 million in 2010 from 1.2 million in 2005. It can be applied in home automation, battlefield
surveillance, health care applications and vehicular environments. A wireless sensor network (WSN)
constitutes a lot of wireless sensing nodes. In addition, a node in WSN consists of one or more sensors,
a radio transceiver, and a microcontroller. The sensor can be used for sensing temperature, pressure,
sound, vibration, motion or position, etc. to collect status from devices or environments. The transceiver
is used to relay the information of the collected status computed by the microcontroller to a center node,
called a gateway or sink. Therefore, a WSN belongs to one type of wireless ad-hoc networks. However,
the nodes in a WSN are usually smaller than that in traditional wireless ad-hoc networks regarding node
size, computing power, memory size, and transmission rage. In other words, the transmission ability,
computing power, and memory size of WSN nodes are limited.
Chapter 6
Wireless Access in Vehicular Environments ......................................................................................... 90
Ching-Wen Chen, Feng Chia University, Taiwan, R.O.C.
The IEEE 1609 standards define communication for wireless access in vehicular environment (WAVE)
services, which enable vehicle-to-vehicle, vehicle-to-roadside, as well as vehicle-to-infrastructure
communications. The standard consists of four parts, which are briefly described in this chapter. IEEE
1609.1 describes the WAVE resource manager which specifies the wireless access method in a WAVE
environment and allows a remote manager application to establish connection with a resource command
processor on an on-board unit. IEEE 1609.2 defines several secure message formats to process mes-
sages for WAVE system. The standard covers methods for securing WAVE management messages and
application messages, which protects messages from attacks such as eavesdropping, spoofing, altera-
tion, replay, and linkable information to unauthorized parties. IEEE 1609.3 defines network services for
WAVE systems, whose network services operate at the network and transport layers of the OSI model
and support both the IPv6 traffics and the WAVE short message services. IEEE 1609.4 describes WAVE
multi-channel operations. It specifies the functions of MAC sublayer management entity and WAVE
MAC with channel coordination. The multi-channel operation provides an efficient mechanism that
controls the operation of upper layer across multiple channels.
Section 3
Chapter 7
Introduction To Global Satellite Positioning System (GPS) ............................................................... 108
Jenq-Muh Hsu, National Chiayi University, Chiayi, Taiwan, R.O.C.
Understanding the right positions and directions of people and objects is a significant issue from the
ancient eras to the present. In the past, people often launched a war in order to satisfy the craving for the
dominating powers and spread their realms. In the recent, Global Satellite Positioning System (GPS) has
become the one of most popular positioning technologies. GPS can provide users precise positioning
information, no matter wherever that may present their own positions. The early GPS positioning technol-
ogy has been widely used in military, marine use, until recently gradually applied into our daily life, e.g.,
automotive navigation, geodesy surveying, etc. In this chapter, the authors will briefly introduce some
GPS issues including the origins of GPS, GPS system architecture, and related GPS applications.
Chapter 8
Vehicle Location and Navigation Systems ......................................................................................... 119
Ben-Jye Chang, National Yunlin University of Science and Technology, Yunlin, Taiwan, R.O.C.
The most driving purpose is to traverse to the destination safely, efficiently, and comfortably. Two types
of approaches could achieve the goals, including the static and dynamic approaches. In the static aspect,
vehicles use the static road and traffic information to navigate. Conversely, in the dynamic aspect, ve-
hicles adopt the dynamic information instead. However, both of the two approaches first require getting
the vehicle’s location and then map the position on an e-map. Thus, this chapter first introduces some
important vehicle location determination algorithms: the dead reckoning and global position system
algorithms, in which the precision of location technologies are compared. Then, the map-matching al-
gorithm is described in detail. Finally, various vehicle navigation approaches are detailed, in which the
important topics include: the navigation architecture, the navigation routing algorithm, and navigation
applications.
Chapter 9
Design and Implementation of Vehicle Navigation Systems .............................................................. 131
Min-Xiou Chen, National Dong-Hwa University, Hualien, Taiwan, R.O.C.
Vehicle Navigation System (VNS) is a complicated and integrated system. A reliable vehicle navigation
system should integrate the wireless communication technologies, positioning technologies, embedded
computer, geographic information database, and so on. The major purpose of the chapter is to help
understanding the architecture of vehicle navigation system. This chapter first introduces the system re-
quirements and system analysis, and show the system platform of vehicle navigation system. The system
platform can be divided into six components. There are the digital map database, positioning devices,
map-matching process, route planning process, route guidance process, human-machine interface, and
wireless communication interface. The design issues and system communication of these components
are detail illustrated in the chapter. Finally, the authors also present some vehicle navigation systems
proposed in the past few years, and show the difference of these systems. The aim of vehicle navigation
system is to guide the vehicle along the optimal path from the starting point to destination. A reliable
vehicle navigation system can reduce the traffic chaos in the city and improve the transportation delay.
In order to achieve reliable vehicle navigation system, the detail system requirements, system analysis,
and system architecture are shown in the chapter. Each component of vehicle navigation system is briefly
illustrated, and the system communication is also described. They also present the architecture of the
proposed vehicle navigation system, and show the difference of these systems. Therefore this chapter
helps understanding the architecture of vehicle navigation system.
Section 4
Chapter 10
Vehicular Metropolitan Area Network Systems Architecture: The WiMAX Network
Reference Model ................................................................................................................................. 144
Cheng Hsuan Cho, National Chung Cheng University, Taiwan, R.O.C.
Jen-Yi Pan, National Chung Cheng University, Taiwan, R.O.C.
The WiMAX NWG develops a network reference model to serve as an architecture framework for
WiMAX deployments and to ensure interoperability among various WiMAX equipment and operators.
The network reference model envisions unified network architecture for supporting fixed, nomadic, and
mobile deployments and is based on an IP service model. We introduce WiMAX network architecture,
WiMAX network entry, mobility management, QoS functional elements, core network planning and
accounting architecture in this section. However, all of them are significant in deploying WiMAX core
network. The operator tries to reach the goals including system performance, reliability, and so on.
On the other hand, the WiMAX operator should consider and balance such many variables in order to
achieve a better situation.
Chapter 11
Interworking of IP Multimedia Subsystem and Vehicular Communication Gateway ........................ 160
Wei-Kuo Chiang, National Chung Cheng University, Chiaya, Taiwan, R.O.C.
An-Nie Ren, National Chung Cheng University, Chiaya, Taiwan, R.O.C.
In recent years, more and more people dream of experiencing various IP-based multimedia application
services when they are driving through their car. However, those multimedia devices in the car may use
different communication protocols such as X.10, Havi, Jini, UPnP and SIP. In order to provide a variety
of IP-based multimedia services to those users in the car, we mainly investigate the issue of interworking
between IP Multimedia Subsystem (IMS) and telematics of the vehicular industry. A service-integrated
platform, Open Service Gateway Initiative Service Platform (OSGi SP), has been proposed to act as a
Residential Gateway (RGW) and to administer the communication between the vehicular environment
and Internet. Besides, a Home IMS Gateway (HIGA), which can be implemented on a NGN RGW,
has been developed by Home Gateway Initiative (HGI) since 2005 to collect the relevant information
of in-car users, devices and services and to manage the IMS sessions for the in-car devices that do not
support IMS functions. With these techniques, the users can enjoy their digital life by interacting with
the home/vehicular network from anywhere.
Section 5
Vehicular Ad Hoc Networks and Delay Tolerant Vehicular Networks
Chapter 12
MAC Protocols in Vehicular Ad Hoc Networks ................................................................................. 183
Chih-Yung Chang, Tamkang University, Taiwan, R.O.C.
With the rapid development of wireless technologies, the Vehicular Ad Hoc Networks (VANETs) have
recently received much attention. VANETs technologies aim to ensure traffic safety for drivers, provide
comfort for passengers and reduce transportation time and fuel consumption with many potential ap-
plications. The achievement of these aims highly relies on efficient MAC protocols which determine the
performance of packet transmission in terms of success rate, delay, throughput and bandwidth utilization.
This chapter reviews the existing MAC protocols developed for VANETs. Initially, the IEEE 802.11p
and DSRC standard are reviewed. Two TDMA-based MAC protocols, called CVIA and VeSOMAC, are
then introduced. In addition, three MAC protocols that cope with the emergency-message broadcasting
problem are proposed. Finally, a reliable MAC protocol which is developed based on the cluster topol-
ogy is reviewed.
Chapter 13
Routing Protocol in Vehicular Ad Hoc Networks ............................................................................... 206
Yuh-Shyan Chen, National Taipei University, Taipei, Taiwan, R.O.C.
Yun-Wei Lin, National Taipei University, Taipei, Taiwan, R.O.C.
Vehicular Ad hoc Network (VANET), a subclass of mobile ad hoc networks (MANETs), is a promising
approach for the intelligent transportation system (ITS). The design of routing protocols in VANETs is
important and necessary issue for support the smart ITS. The key difference of VANET and MANET
is the special mobility pattern and rapidly changeable topology. It is not effectively applied the existing
routing protocols of MANETs into VANETs. In this chapter, the authors mainly survey new routing
results in VANET. They introduce unicast protocol, multicast protocol, geocast protocol, mobicast pro-
tocol, and broadcast protocol. It is observed that carry-and-forward is the new and key consideration for
designing all routing protocols in VANETs. With the consideration of multi-hop forwarding and carry-
and-forward techniques, min-delay and delay-bounded routing protocols for VANETs are discussed in
VANETs. Besides, the temporary network fragmentation problem and the broadcast storm problem are
further considered for designing routing protocols in VANETs. The temporary network fragmentation
problem caused by rapidly changeable topology influence on the performance of data transmissions.
The broadcast storm problem seriously affects the successful rate of message delivery in VANETs. The
key challenge is to overcome these problems to provide routing protocols with the low communication
delay, the low communication overhead, and the low time complexity.
Chapter 14
Applications in Vehicular Ad Hoc Networks ...................................................................................... 229
Hua-Wen Tsai, National University of Tainan, Tainan, Taiwan, R.O.C.
Yi-Shiang Chang, National University of Tainan, Tainan, Taiwan, R.O.C.
The various sensors and wireless communication devices have been extensively applied to daily life
due to the advancements of microelectronics mechanism and wireless technologies. Recently, vehicular
communication systems and applications become more and more important to people in daily life. Ve-
hicular communication systems that can transmit and receive information to and from individual vehicles
have the potential to significantly increase the safety of vehicular transportation, improve traffic flow on
congested roads, and decrease the number of people of deaths and injuries in vehicular collisions effec-
tively. This system relies on direct communication between vehicles to satisfy the communication needs
of a large class of applications, such as collision avoidance, passing assistance, platooning. In addition,
vehicular communication systems can be supplemented by roadside infrastructure to access Internet and
other applications. This system forms a special case of mobile ad hoc networks called Vehicle Ad Hoc
Networks (VANETs). They can be formed between vehicles with vehicle to vehicle (V2V) communica-
tion or between vehicles and an infrastructure with vehicle to infrastructure (V2I) communication. The
applications and characteristics of VANETs are introduced and presented in this Chapter.
Chapter 15
DTN Technologies for Vehicular Networks........................................................................................ 252
A Delay Tolerant Network (DTN) is one type of challenged network where network contacts are inter-
mittent or link performance is highly variable or extreme. In such a network, a complete path does not
exist from source to destination for most of the time. In addition, the path can be highly unstable and
may change or break unexpectedly. To make communication possible in a delay tolerant network, the
intermediate nodes need to take custody of data during the blackout and forward it toward the destina-
tion when the connectivity resumes. A vehicular network nicely falls into the context of DTN since the
mobility of vehicles constantly causes the disruption of link connectivity’s between vehicles. In this
chapter, the authors discuss some research challenges and issues which might occur in a Delay Tolerant
Network and how they are related to vehicular networks.
Section 6
Chapter 16
Simple Transporation Management Framework................................................................................. 271
The Simple Transportation Management Framework (STMF) specifies a set of rules and protocols which
can be used to organize, describe, and exchange transportation management information between trans-
portation management applications and equipments. The STMF framework consists of four elements,
including Management Information Base (MIB), Structure and Identification of Management Informa-
tion (SMI), Simple Network Management Protocol (SNMP), and Simple Transportation Management
Protocol (STMP). MIB is a collection of management objects written in ASN.1 notation. SMI is the
definition of how to create management objects and a hierarchical definition of nodes where management
objects will be attached for unique identification. SNMP is a communications protocol for configuring
and monitoring of network devices. STMP is a variation of SNMP to address low-bandwidth commu-
nication links and real-time device monitoring.
Chapter 17
Vehicular System Management Architecture and Application Platform ............................................ 290
Notably, not all telematics services can be used in telematics terminals as a result of the varied platform
standards. The main issues are that most telematics technologies depend on vertical, proprietary and
closed per-OEM Original Equipment Manufacture (OEM) platforms, forming islands of non-interop-
erable technology and preventing third-party service providers from creating valuable services. In this
study, the Open Gateway Service Initiative Vehicle Expert Group (OSGi/VEG) was integrated into
an Android platform to generate a vehicular Android/OSGi platform that has the advantages of both
original platforms, such as remote management, rich class sharing, proprietary vehicular applications,
security policies, easy management of application programming interface (APIs), and an environment
with increased openness. Furthermore, this study integrates the cloud computing mechanism into the
Android/OSGi platform, which allows service providers to upload their telematics bundles onto storage
clouds via the provisioning server.
Chapter 18
Remote Vehicular System Management Functions and Information Structure .................................. 310
Due to the rapid development of information technology, the network has already spread to every corner of
vehicle. With all kinds of ECU devices appear in the vehicle, and it brings the more and more convenient
living. On purpose solving heterogamous technologies that are incompatible with each other, developed
a “WBEM-based Remote Management and Heterogeneous Vehicular Network Diagnosis System” on
OSGi Gateway. This system can focus on a variety of problems come from vehicle network, and find
out what are the problems or where are the problems happened. If the problem still can not be solved
properly, we must to seek for help from remote managers. The users can acquire enough information
without understanding how to control every device, so that the users can help near diagnosis system to
solve vehicle network’s problems and to promote the abilities of near network diagnosis.
Chapter 19
Using Wireless Mesh Network for Traffic Control ............................................................................. 331
Wireless mesh networks (WMN) have attracted considerable interest in recent years as a convenient,
flexible and low-cost alternative to wired communication infrastructures in many contexts. However,
the great majority of research on metropolitan-scale WMN has been centered around maximization
of available bandwidth, suitable for non-real-time applications such as Internet access for the general
public. On the other hand, the suitability of WMN for missioncritical infrastructure applications remains
by and large unknown, as protocols typically employed in WMN are, for the most part, not designed for
real-time communications. In this chapter, the authors describe a real-world testbed, which sets a goal
of designing a wireless mesh network architecture to solve the communication needs of the traffic con-
trol system in Sydney, Australia. This system, known as SCATS (Sydney Coordinated Adaptive Traffic
System) and used in over 100 cities around the world, connects a hierarchy of several thousand devices
-- from individual traffic light controllers to regional computers and the central Traffic Management
Centre (TMC) - and places stringent requirements on the reliability and latency of the data exchanges.
The authors discuss some issues in the deployment of this testbed consisting of 7 mesh nodes placed at
intersections with traffic lights, and show some results from the testbed measurements.
Section 7
Mobility Model, Simulation, and Security
Chapter 20
Mobility Models of Vehicular Networks ............................................................................................ 348
A key component for VANET simulations is a realistic vehicular mobility model that ensures conclusions
drawn from simulation experiments will carry through to real deployments. However, VANET simulations
raise many new questions about suitable levels of details in simulation models. To get accurate results,
the mobility models of Vehicular Networks should be as realistic as possible, and involve road-maps
with all constraints and facilities related to the vehicular movement. In this chapter, the authors provide
an overview of some mobility models that are relevant to VANETs. The criteria of applicability they
consider here is the employment of road maps, and thus limiting the nodes movements into the routes,
instead of moving them in a wide open area. They compare different models based on the parameters
they use. For instance, some models use traffic control mechanisms (stop signs or traffic lights) at route
intersections, and some just assume continuous movement at these points. Some assume routes to be
single-lane, some others support multi-lanes routes. Some define the security distance, while others just
ignore this parameter.
Chapter 21
MOVE: A Practical Simulator for Mobility Model in VANET .......................................................... 355
Vehicular Ad-Hoc Network (VANET) is surging in popularity, in which vehicles constitute the mobile
nodes in the network. Due to the prohibitive cost of deploying and implementing such a system in real
world, most research in VANET relies on simulations for evaluation. A key component for VANET simula-
tions is a realistic vehicular mobility model that ensures conclusions drawn from simulation experiments
will carry through to real deployments. However, VANET simulations raise many new questions about
suitable levels of details in simulation models for nodes mobility. In VANET simulations, the mobility
models used affect strongly the simulation output. The researchers need to decide what level of details
are required for their simulations. In this chapter, the authors introduce a tool MOVE that allows users
to rapidly generate realistic mobility models for VANET simulations. MOVE is built on top of an open
source micro-traffic simulator SUMO. The output of MOVE is a realistic mobility model and can be
immediately used by popular network simulators such as ns-2 and Qualnet. They show that the simula-
tion results obtained when using a realistic mobility model such as MOVE are significantly different
from results based on the commonly used random waypoint model. In addition, they evaluate the effects
of details of mobility models in three case studies of VANET simulations (specifically, the existence of
traffic lights, driver route choice and car overtaking behavior) and show that selecting sufficient level
of details in the simulation is critical for VANET protocol design.
Chapter 22
Security Attacks of Vehicular Networks ............................................................................................. 369
Jen-Chun Chang, National Taipei University, Taiwan, R.O.C.
Chun-I Fan, National Sun Yat-sen University, Taiwan, R.O.C.
Ruei-Hau Hsu, National Sun Yat-sen University, Taiwan, R.O.C.
The application of vehicular ad hoc network (VANET) improves driving safety and traffic management.
Due to the above applications, security attacks on VANET can be serious threats all the time. VANET is a
special form of mobile ad hoc network (MANET). Hence any attacks exist on MANET also can be arisen
on VANET. Moreover, some special attacks can be raised on VANET, which don’t exist on MANET.
Nevertheless, some characteristics of VANET can be positive effects and some can be negative effects
on security issues. Before designing the security mechanism to defend attacks, the authors should take
the positive effects and avoid the negative effects on the security of VANET. Furthermore, they class all
possible attacks of VANET from every network layer. The authors also introduce the reason of forming
every attack and the possible effect on VANET in detail. Therefore this chapter helps understanding the
latent threats and the useful resources of security issues on VANET.
Compilation of References ............................................................................................................... 380
About the Contributors .................................................................................................................... 398
Index ................................................................................................................................................... 405

xix
Foreword
The Human Resource Program for Information and Communication Technology sponsored by the
Ministry of Education (MOE) in Taiwan, lead by myself, has been a key program in the past 10 years
to bridge possible gaps between university education and industrial human resource demands, and to
train and cultivate sufficient as well as high-quality and skilled young professionals in the blooming
global telecommunication market. After a detailed analysis of the ICT industrial trend by a review board
consisted of experts and professionals from both academic and industrial sector around 2006, MOE
in Taiwan decided to focuses the talent cultivation direction only on a few selected areas of emerging
technology. Telematics and Vehicular Networks has emerged as one such important technology since
the research progress in the telematics and vehicular networking was so significant that related industry
has become booming around the world and one can easily expect future drivers can enjoy the benefit
from related smart telematics products and services. We also believe telematics and vehicular networks
can be categorized as a special kind of Green ICT Technologies since it also helps to save energy in
many scenarios.
In the past few decades, Taiwan’s telecommunication and ICT industry has experienced a long pe-
riod of high growth and fast technology evolution. For example, the communication industry in Taiwan
has increased up to 8 folds around 10 years, with its manufacturing capacity ranging from traditional
LAN switches to 3G smart phones. Without surprises, recent industry trends in smart cars and vehicu-
lar networks have also created a strong demand on talent engineers with good hand-on experiences in
related products and services. Again, universities and the academic community have been asked to keep
upgrading key ICT courses and laboratories to link up with the telematics industry in a timely fashion.
In response to this demand, a prestigious team was selected from the academic community in Taiwan
in 2007, to aggregate teaching resources, refine the essential courseware, and enhance experiment en-
vironments for training talented students in this field. An intercollegiate telematics promotion center
was also established for completing this task and Professor Chung-Ming Huang, National Cheng-Kung
University, was selected to lead this center for his dedication and the knowledge and research experience
he has accumulated in this field.
This book, Telematics Communication Technologies and Vehicular Networks: Wireless Architectures
and Applications, edited by Prof. Chung-Ming Huang and Prof. Yuh-Shyan Chen, is a work contributed
by such a group of telematics experts and professors in this field. This book has successfully covered a
wide range of technical topics, including vehicular network architecture, related communication proto-
cols, ITS/telematics applications, navigation systems, location based services and embedded systems.
xx
Many chapters in this book is self-guided and can be used a tutorial. In general, it should be a valuable
textbook, guidance and/or reference for students, researchers, engineers, technical managers, and other
professionals in this field. I believe all readers can enjoy reading this book.
Zsehong Tsai, Professor

Human Resource Program Office for ICT
National Taiwan University, Taiwan, R.O.C.
xxi
Preface
TelemaTics communicaTion Technologies and Vehicular

neTworks: wireless archiTecTures and applicaTions
Telematics communication technologies and vehicular networks have been identified as key technologies
for increasing road safety and transport efficiency. Telematics communication technologies and vehicular
networks aim to ensure traffic safety for drivers, provide comfort for passengers and reduce transportation
time and fuel consumption. The development of vehicular communication and networking technologies
are expected to enable many potential applications, including automatic collision notification and pre-
vention, emergency management, assistances for safe driving, real-time traffic congestion notification,
location-based driver information services, high-speed tolling, vehicle tracking, automobile Internet
access, and many others. To facilitate these applications, many different new types of communication
and networking would be involved, including intra-vehicle, vehicle-to-vehicle, vehicle-to-roadside and
vehicle-to-infrastructure communications. This book aims to provide a fast and complete view of all
aspects related to telematics communication technologies and vehicular networks.
This book is intended for graduate students, researchers, engineer, and practitioners who are interested
in acquiring a global view of the set of techniques and protocols that have been referred to as “Telemat-
ics communication technologies and vehicular networks: wireless architectures and applications” in the
literature. The book can serve as a reference resource for researchers, engineers, and developers working
in the field of telematics technologies.
This book includes 22 chapters which are classified into 7 Sections. Section 1 introduces the vehicu-
lar networks and intelligent transportation systems. Section 2 describes embedded system architecture
and communication protocols. Section 3 reports location based services. Section 4 provides integrated
vehicular applications. Section 5 presents vehicular ad hoc networks and delay tolerant vehicular net-
works. Section 6 explains management and traffic control. Finally, Section 7 talks about mobility model,
simulation, and security for telematics communication technologies and vehicular networks.
The first section of the book, Introduction of Vehicular Networks and Intelligent Transportation Sys-
tems, presents introductory materials that is preparatory for what us described in the rest of the book.
Chapter 1, by Ming-Chiao Chen and Teng-Wen Chang, gives a short introduction to vehicular net-
work architectures. This chapter discusses the architectures in the vehicular network environment, and
introduces the overview of in-vehicle and out-vehicle network architectures. This chapter also discusses
the next-generation vehicular network architecture, the modern in-vehicle networks, on-board computers
and the Internet, mobile telecommunications and telematics applications in the ground vehicles.
Chapter 2, by Yao-Chung Chang, introduces vehicular network applications. Four applications of the
vehicular network are surveyed and discussed in this chapter. They are vehicular network application
xxii
services, vehicular network application managements, the platform technologies of vehicular network
application, and the future vehicular network application and deployment.
Chapter 3, by Da-Jie Lin and Chyi-Ren Dow, explains intelligent transportation systems. Intelligent
Transportation Systems (ITS) combines high technology and improvements in information systems,
communication, sensors, and relevant mathematical methods with the conventional world of surface
transportation infrastructure to increase the capacity of transportation systems and to improve the level
of services. There are four major goals of ITS, including safety, environmental protection, efficiency,
and economy.
The second section of the book, Embedded System Architecture and Communication Protocols, pres-
ents the vehicular embedded system architecture, data communication protocols for vehicular network,
and wireless access techniques in vehicular environments.
Chapter 4, by Chung-Ping Young, explains the vehicular embedded system architecture. The car elec-
tronics plays an increasingly important role in automobile industry. The embedded system has already
been extensively employed for improving the operation and performance of vehicles, such as safety,
comfort, convenience, and environmental protection. In terms of electronic system, an automobile is a
distributed embedded system, and the control messages to each electronic control unit (ECU), go through
the in-vehicle network. An ECU is an embedded processor or computing system, integrated with a data
acquisition device or an electromechanical driver.
Chapter 5, Cheng-Min Lin, and Tzong-Jye Liu, reports the data communications inside vehicular
environments. ZigBee is based on IEEE 802.15.4 which specifies the physical layer and medium access
control (MAC) for low-cost and low-power LR-WPAN. ZigBee is successfully produced into a low-cost
controller applied for automotive applications, including vehicle control and status monitoring.
Chapter 6, by Tzong-Jye Liu and Ching-Wen Chen, gives the wireless access in vehicular environments.
This chapter describes the IEEE 1609 standard for Wireless Access in Vehicular Environment (WAVE)
services, which enable vehicle-to-vehicle, vehicle-to-roadside, as well as vehicle-to-infrastructure com-
munications. The standard consists of four parts, IEEE 1609.1, IEEE 1609.2, IEEE 1609.3, and IEEE
1609.4, which are described in this chapter.
The third section of the book, Location Based Services, provides the useful knowledge and tech-
nique of location based services for supporting telematics communication technologies and vehicular
networks.
Chapter 7, by Jenq-Muh Hsu, provides the introduction of GPS. Global Satellite Positioning Sys-
tem (GPS) recently has become the one of most popular positioning technologies. This chapter briefly
introduces some GPS issues including the origins of GPS, GPS system architecture, and related GPS
applications.
Chapter 8, by Ben-Jye Chang, presents the vehicle location and navigation systems. This chapter
introduces important vehicle location determination algorithms: the dead reckoning and global position
system algorithms, in which the precision of location technologies are compared. The map-matching
algorithm is then described. Various vehicle navigation approaches are explained.
Chapter 9, by Min-Xiou Chen, discusses the design and implementation of vehicle navigation sys-
tems. The major purpose of the chapter is to understand the architecture of vehicle navigation system.
This chapter introduces the system requirements and system analysis, and show the system platform
of vehicle navigation system.
In the fourth section of the book, Integrated Vehicular Application, presents the vehicular metropoli-
tan area network systems architecture and the interworking of IP multimedia subsystem and vehicular
communication gateway.
xxiii
Chapter 10, by Cheng Hsuan Cho and Jen-Yi Pan, reports the vehicular metropolitan area network
systems architecture: the WiMAX network reference model. This chapter introduces WiMAX network
architecture, WiMAX network entry, mobility management, QoS functional elements, core network plan-
ning and accounting architecture. WiMAX technique is the one of important wireless access techniques
for the vehicular communication.
Chapter 11, by Wei-Kuo Chiang and An-Nie Ren, presents the interworking of IP multimedia subsystem
and vehicular communication gateway. To provide a variety of IP-based multimedia services to those
users in the car, this chapter investigates the issue of interworking between IP Multimedia Subsystem
(IMS) and telematics of the vehicular industry.
Section 5 of the book, Vehicular Ad Hoc Networks and Delay Tolerant Vehicular Networks, presents
MAC, routing protocols, and applications for vehicular ad hoc networks, and DTN technologies for
vehicular networks.
Chapter 12, by Chih-Yung Chang, presents MAC protocols in vehicular ad hoc networks. The
achievement of requirements of VANETs highly relies on efficient MAC protocols which determine the
performance of packet transmission in terms of success rate, delay, throughput and bandwidth utilization.
This chapter reviews existing MAC protocols developed for VANETs.
Chapter 13, by Yuh-Shyan Chen and Yun-Wei Lin, presents routing protocols in vehicular ad hoc
networks. The design of routing protocols in VANETs is important and necessary issue for support the
smart ITS. This chapter discusses existing routing results, including unicast protocol, multicast protocol,
geocast protocol, mobicast protocol, and broadcast protocol, in VANETs.
Chapter 14, by Tzung-Shi Chen, Hua-Wen Tsai, and Yi-Shiang Chang, discusses applications in
vehicular ad hoc networks. Vehicular communication systems and applications become more and more
important to people in daily life. The applications of VANETs are introduced in this Chapter.
Chapter 15, by Kun-Chan Lan, reports DTN technologies for vehicular networks. A Delay Toler-
ant Network (DTN) is one type of challenged network where network contacts are intermittent or link
performance is highly variable or extreme. This chapter discusses some research challenges and issues
which might occur in a Delay Tolerant Network.
Section 6 of the book, Management and Traffic Control, presents the simple transportation manage-
ment framework, the vehicular system management architecture and application platform, the remote
vehicular system management function and information structure, and the wireless mesh network for
the traffic control.
Chapter 16, by Chyi-Ren Dow, discusses the simple transportation management framework. The
Simple Transportation Management Framework (STMF) specifies a set of rules and protocols which can
be used to organize, describe, and exchange transportation management information between transporta-
tion management applications and equipments.
Chapter 17, by Teng-Wen Chang and Jiann-Liang Chen, reports the vehicular system management
architecture and application platform. In this chapter, the Open Gateway Service Initiative Vehicle
Expert Group (OSGi/VEG) was integrated into an Android platform to generate a vehicular Android/
OSGi platform.
Chapter 18, by Teng-Wen Chang and Jiann-Liang Chen, presents the remote vehicular system man-
agement functions and information structure. On purpose solving heterogamous technologies that are
incompatible with each other, this chapter develops a “WBEM-based Remote Management and Hetero-
geneous Vehicular Network Diagnosis System” on the OSGi Gateway.
Chapter 19, by Kun-Chan Lan, presents the wireless mesh network for traffic control. Wireless mesh
networks (WMN) have attracted considerable interest in recent years. This chapter described a real-world
xxiv
testbed, which sets a goal of designing a wireless mesh network architecture to solve the communication
needs of the traffic control system in Sydney, Australia.
The final section of the book, Mobility Model, Simulation, and Security, provides a detailed de-
scription of mobility models of vehicular networks, REALISTIC simulation of vehicular networks, and
security attacks of vehicular networks.
Chapter 20, by Kun-Chan Lan, reports the mobility models of vehicular networks. A key component
for vehicular network simulation is a realistic vehicular mobility model that ensures conclusions drawn
from simulation experiments will carry through to real deployments. To get accurate results, the mobility
models of vehicular networks should be as realistic as possible, and involve road-maps with all constraints
and facilities related to the vehicular movement. Therefore, this chapter provides an overview of some
mobility models that are relevant to vehicular networks.
Chapter 21, by Kun-Chan Lan, presents the realistic simulation of vehicular networks. This chapter
introduces a tool, MOVE, that allows users to rapidly generate realistic mobility models for vehicular
network simulations. MOVE is built on top of an open source micro-traffic simulator SUMO. The out-
put of MOVE is a realistic mobility model and can be immediately used by popular network simulators
such as ns-2 and Qualnet.
Chapter 22, by Jen-Chun Chang, Chun-I Fan, and Ruei-Hau Hsu, discusses the security attacks of
vehicular networks. This chapter classifies all possible attacks of vehicular network from every net-
work layer, and also introduces the reason of forming every attack and the possible effect on vehicular
networks.
Finally, we thank all contributors of the book for their outstanding contributions. We hope you
will enjoy reading this book as we did and you will find this issue informative and helpful in keeping
yourselves up-to-date in the fast changing field of telematics communication technologies and vehicular
networks, from wireless architectures to applications
Chung-Ming Huang
National Cheng Kung University, Taiwan, R.O.C.
Yuh-Shyan Chen
National Taipei University, Taiwan, R.O.C.
July 2009
xxv
Acknowledgment
The editors and authors are working members of The Promotion Center for Telematics Consortium
(PCTC), which is part of the Information and Communication Human Resource Program, Ministry of
Education (MOE), Taiwan, R.O.C. Thanks for the program of MOE such that the editors and authors
can be grouped together. Many thanks to all authors for their hard work and cooperation for delivering
their chapters. We also would like to thank Joel A. Gamon of IGI Global for his help and encouragement
during this period.
Section 1
Introduction of Vehicular
Networks and Intelligent
Transporation Systems
1
Chapter 1
Network Architectures
Ming-Chiao Chen
National Taitung University, Taitung, Taiwan, R.O.C.
Teng-Wen Chang
National Taiwan University, Taipei, Taiwan, R.O.C.
absTracT
A vehicular network organizes and connects vehicles with each other, and with mobile and fixed-locations
resources. This chapter discusses the architectures in the vehicular network environment. The authors
introduce the overview of in-vehicle and out-vehicle network architectures. An automobile in an in-vehicle
network adopts four vehicle bus protocols, CAN (Controller Area Network), LIN (Local Interconnect
Network), MOST (Media Oriented Systems Transport) and FlexRay. However, these protocols cannot
intercommunicate with each other. Therefore, the OSEK operating system is designed as standard software
architecture for the various ECUs (Electronic Control Units). In the out-vehicle network, the OBU (On
Board Unit) in the automobile can communicate with the infrastructure via the Internet. The authors
discuss next-generation vehicular network architecture, the modern in-vehicle networks, on-board com-
puters and the Internet, mobile telecommunications and telematics applications in the ground vehicles,
and finally, we introduce future desired features. This chapter discusses the architectures in vehicular
network environment. The first section introduces the overview of in-vehicle and out-vehicle network
architectures. The next section describes in-vehicle network architecture for disaster communication
network by combining various automotive bus protocols. The third section describes the out-vehicle
network architecture for disaster communication network by combining various wireless LANs. The
last section discusses next-generation vehicular network architecture, the modern in-vehicle networks,
on-board computers and the Internet, mobile telecommunications and telematics applications in the
ground vehicles, and introduces future desired features.
DOI: 10.4018/978-1-60566-840-6.ch001
Copyright © 2010, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
Introduction of Vehicular Network Architectures
inTroducTion by the convention center even before stepping

out of the car.
A vehicular network organizes and connects Current telematics systems depend on mobile
vehicles with each other, and with mobile and infrastructures to deliver telematics services to
fixed-locations resources (Wu et al., 2005). service users. Therefore, deploying a telematics
Many telematics architectures, including navi- service between mobile networks it is a very
gation services architecture, traffic information expensive task. System developers need to have
architecture, location-based services architecture, strong knowledge about the underlying mobile
entertainment services architecture, emergency network. Additionally, a telematics terminal
and safety services architecture have been pro- cannot be applied for the telematics service
vided. In these architectures, traffic information from another telematics service provider, since
and navigation services are generally provided telematics service developers devise their own
by central TSPs (Telematics Service Providers). protocols between telematics terminals and a
Emergency and safety services are supplied by an service provider.
on-board platform, which is likely to be installed Figure 1 shows an overview of in-vehicle
by the individual car manufacturers. network architecture and out-vehicle network
In contrast to these conventionally adopted architecture. An automobile in an in-vehicle
architectures, telematics architectures are rarely network adopts four vehicle bus protocols, CAN
applied in public local hotspots such as public (Controller Area Network), LIN (Local Intercon-
parking lots, hotels, restaurants, airports and nect Network), MOST (Media Oriented Systems
shopping centers. In local hot-spot architecture, Transport) and FlexRay. However, these protocols
a vehicle is considered as an alternative mobile cannot intercommunicate with each other. There-
computing platform (logically equivalent to a fore, the OSEK operating system was designed
PDA, laptop or a cellular phone) with short-range as standard software architecture for the various
localized WLAN (Wireless LAN) devices such ECUs (Electronic Control Units). In the out-
as Bluetooth and Wi-Fi. This local (hot-spot) vehicle network, the OBU (On Board Unit) in the
architecture allows the car driver to interact with automobile can communicate with the infrastruc-
many local services. Telematics architectures ture via the Internet. The remote home service
will be useful for telematics services only for and remote vehicular service providers provide
vehicles providing traditional services such as particular services to an automotive user. The in-
traffic information, navigation services provided vehicle and out-vehicle network architectures are
by the central TSP, and local services provided discussed in detail in the next two sections.
by distributed third-party service providers that
can supply the appropriate contextual data. For
instance, consider a driver wishing to drive a car in-Vehicle neTwork
to a convention center. The driver initially finds archiTecTure
routes to the center using a navigator, then selects
a route free from traffic jam based on traffic This section introduces an in-vehicle network
information from a TSP. The car automatically architecture for disaster communication network
discovers the resources of the convention center, that combine different automotive bus protocols,
obtains directions to the designated parking lot, namely Controller Area Network (CAN), Local
and makes associated payments using the WLAN Interconnect Network (LIN) and the recently
communication, as it enters the premises. The developed FlexRay protocol standard. Moreover,
driver can obtain various local services provided the OSEK/VDX operating system, a joint project
2
Figure 1. In-vehicle and out-vehicle network architecture
of the automotive industry, manipulates automo- message format as it is defined in CAN specifica-
tive messages between different bus protocols to tion 1.2, and part B describes both standard and
support efficient usage of resources for automotive extended message formats. To achieve design
control unit application software. transparency and implementation flexibility, CAN
is divided into three layers, the object, transfer
controller area network and physical layers. Figure 2 Johansson et al.
(2003) illustrates a CAN bus with three nodes.
German automotive system supplier Robert Bosch Johansson et al. (2003) described an application
created CAN in the mid-1980s for automotive ap- process example of a node. A temperature sensor
plications as an effective means of allowing robust decides when to request the transmission of a mes-
serial communication (Pazul et al., 1999). The goal sage frame. The frame consists of a data field and
was to establish a standard for more reliable and overhead, such as the identifier and control fields.
efficient communication by integrating devices, Since the application processes are asynchronous,
sensors and actuators in a system for real-time the bus has a mechanism called CSMA/CD, car-
control applications. The CAN protocol combines rier sense multiple access/collision avoidance,
networks and electronic control units thus reduc- for resolving conflicts. The protocol listens to the
ing both wiring harness and complexity. CAN network in order to avoid collision.
has now gained widespread use in automotives CAN has four frame types, namely data, re-
and mobile applications, as well as in industrial mote, error and overload frames. The data frame
automation applications. is the only frame that is adopted for data transmis-
The CAN (Robert Bosch et al., 1991) speci- sion. The data frame has two message formats,
fication has two parts. Part A describes the CAN base frame format (with 11-bit identifier) and
3
Figure 2. Three nodes connected through a CAN

sub-network of a CAN bus to integrate intelligent
bus
sensor devices or actuators in modern cars.
The main features of LIN include:
• Mono-master, up to 15 slaves
• 1 wired bus
• Bitrates 1-20 Kbits/s: 2.4Kbits, 9.6Kbits
and 19.2Kbits are usually used in automo-
extended frame format (with 29-bit identifier). tive applications
This discussion focuses on the base frame format, • Multicast (broadcast) messages
shown in Figure 3 Johansson et al. (2003). The • Self-synchronization of the slave (only the
start-of-frame (SOF) bit denotes the start of the master has an accurate clock as crystal)
frame transmission. It is followed by the 11-bit • Messages with 2,4 or 8 data bytes, and 3
identifier and the remote transmission request control bytes
(RTR) bit. The control field consists of 6 bits, and • Error detection by 8 bits checksum and 2
denotes the number of bytes of data that follow in parity bits in identifier
the data field. The data field contains 0-8 bytes. It • Physical layer: ISO9141
is followed by the cyclic redundancy checksum • Sleep / wake-up capability
(CRC) field, which enables the receiver to check
whether the received bit sequence is corrupt. The The LIN implements serial communication in
transmitter adopts 2-bit acknowledgment (ACK) a state-machine, with small microcontrollers or
field, ACK slot bit, to receive an acknowledgment CPLDs. A slave ECU does not need an accurate
of a valid frame from any receiver. The end of a clock, so can replace crystals or resonators with an
message frame is a 7-bit end-of-frame (EOF). RC cell. This is a cost-effective means of designing
smart actuators or sensors, or smart connectors.
local interconnect network The specification describes three of the seven
layers of the OSI model namely physical layer,
The Local Interconnect Network Bus (LIN-Bus) data link layer and application layer.
is a vehicle bus standard or computer network- A LIN network consists of a LIN master and one
ing bus-system used within current automotive or more LIN slaves. The LIN bus in an automotive
network architectures. The LIN specification is application is generally connected between smart
enforced by the LIN-consortium, with the first sensors or actuators and an Electronic Control
exploited version 1.1, released in 1999. The speci- Unit (ECU), which is often a gateway with a
fication has since evolved to version 2.1 to satisfy CAN bus. A LIN network may have several LIN
current networking needs. The LIN bus is a small buses with no interconnection between them, as
and slow network system that is used as a cheap shown in Figure 4. This network markedly dif-
Figure 3.CAN message frame
4
Figure 4. LIN bus
fers from other low-cost buses as K-line which in-vehicle networking, and therefore does not
was intended to link all the ECUs to an external replace existing networks, but instead works in
analysis approach for diagnosis purposes. conjunction with already well-established sys-
tems, such as the controller area network (CAN),
Flexray local interconnect network (LIN) and media ori-
ented systems transport (MOST). In Figure 5, an
FlexRay is a new communication standard that in-vehicle network with FlexRay serving as the
provides a high-speed serial communication, time- backbone provides determinism for engine control
triggered bus and fault-tolerant communication and fault tolerance for steer-by-wire, brake-by-
between electronic devices for future automotive wire and other advanced safety applications.
applications (FlexRay, 2005; Xu et al., 2008).
FlexRay (2005) was developed for the next gen- vasos and osek/VdX
eration of automobiles and future applications, in-Vehicle management system
including x-by-wire, by a consortium founded
by BMW, Bosch, DaimlerChrysler and Philips vASOS (Sun et al., 2006) (Vehicular Applica-
in 2000 (FlexRay Consortium, 2009). tion Specific Embedded Operating Systems)
The FlexRay protocol provides a high-speed, is designed specifically for vehicle use, and is
deterministic and fault-tolerant communications designed to run on a high-performance user in-
technology. FlexRay is designed specifically for terface computer. It fulfills and provides specific
Figure 5. FlexRay backbone
5
device drivers, such as CAN/LIN buses, which • Plug and play device driver interfaces for
are used to communicate with other ECU nodes expansibility.
for diagnostic purposes, and other fundamental • Emphasis on network control methods, es-
network functions for the vehicle domain. pecially wireless networks.
OSEK/VDX (Kim et al., 2007) (Open systems • Fast boot-up and good power management.
and corresponding interfaces for automotive
electronics / Vehicle Distributed eXecutive) is an Figure 6 shows the OSEK/VDX In-vehicle
open vehicular industry standard that was founded Management System, which consists of three
as a French-German joint project, and which is components:
now drawing attention world-wide. The primary
aims of OSEK/VDX are to address the high cost • vASOS COM: Although OSEK/VDX
in developing and redeveloping Electronic Control COM provides a rich set of communica-
Unit (ECU) software, and to improve the compat- tion facilities, many applications, such as
ibility of those applications. The most important this vASOS prototype model, probably
advantages of OSEK/VDX include portability and only require a minimum subset of this
reliability. OSEK/VDX includes specifications for functionality and all observe OSEK/VDX
OSEK Operating System (OS), OSEK Commu- COM specification.
nication (COM), OSEKNetwork Manager (NM) • OSEK/VDX COM module is composed
and OSEK Implementation Language (OIL). of:
The properties of vASOS are as follows: ◦ An Interaction layer: The layer that
provides communication services for
• Focusing on system real-time performance, the transfer of application messages.
scalability and robustness. ◦ A Network layer: The layer that
• Small kernel size, small memory footprint, provides services for the unacknowl-
low-cost and high efficiency. edged and segmented transfer of
Figure 6. OSEK/VDX in-vehicle network management architecture
6
application messages. The network by combining different wireless networks. Cur-

layer provides flow control mecha- rently available wireless technologies, including
nisms to allow interfacing of commu- IEEE802.11p (2007), IEEE802.16 (WiMAX)
nication peers featuring different lev- (2006), and cellular networks, are combined with
el of performance and capabilities. a mobile router, and loaded on a car to construct
◦ A Data link layer interface that pro- a mobile network node.
vides services for the unacknowl-
edged transfer of individual data ieee 802.11p
packets over a network to the layers
above it. A vehicular environment requires a set of new
• vASOS NM: The NM module focuses requirements on modern wireless communica-
mainly on ensuring the safety and the re- tion systems. Vehicular safety communications
liability of a communication network for applications cannot tolerate long connection
ECUs. This module provides two alterna- establishment delays before communicating with
tive mechanisms for network monitoring: other vehicles encountered on the road.
indirect monitoring by monitored appli- The IEEE 802.11p standard, also referred to
cation messages and direct monitoring by as Wireless Access for the Vehicular Environ-
dedicated NM communication using the ment (WAVE), is designed to solve these issues.
token principle. The WAVE protocol provides improvements to
the physical (PHY) and medium access control
ouT-Vehicle neTwork (MAC) layers of the existing 802.11 wireless
archiTecTure standards.
The Department of Transportation of the
This section introduces out-vehicular network United States advocates and supports ITS (Intel-
architecture for disaster communication network ligent Transportation System) based on dedicated
Figure 7. Vehicle safety communication examples
7
short range communications (DSRC (Xiang et al., is an optional element depending on deployment
2006)) systems that provide vehicle-to-vehicle and choices. In the simplified form (Simple IP), the
vehicle-to-roadside information exchanges. ITS user traffic bypasses the HA in the CSN.
largely focuses on enabling public safety applica- The user traffic is tunneled as payload between
tions that can save lives and improve traffic flow. the BS and the ASN-GW. The Proxy Mobile IP
Jiang et al. (2006) Two such application scenarios protocol handles mobility between the ASN-GW
are shown in Figure 7. Private services are also al- and the HA. A WiMAX BS can potentially connect
lowed in order to spread the deployment costs, and to any ASN-GW that it can reach via IP connectivity
to encourage the quick development and adoption (flex R6). This flexibility helps decrease mobility-
of DSRC technologies and applications. related signaling in the network, since the same
The DSRC spectrum, as shown in Figure 8 ASN-GW can serve the user’s active IP session
(Eichler et al., 2007; Jiang et al., 2006), is struc- while the user is moving across several different BSs
tured into seven channels, each with a bandwidth of (e.g., ASN-GW relocation is rarely required). The
10-MHz. The control channel (CCH) is restricted R8 interface can facilitate the context transfer and
to safety communications only. The two channels handover optimization when the user moves from
at the ends of the spectrum band are reserved for one BS to another. Figure 9 shows the end-to-end
special applications. The service channels (SCH) is Mobile WiMAX network architecture.
available for both safety and non-safety usage. The WiMAX Forum has defined an architecture
that determines how a WiMAX network connects
wimaX with other networks, and a variety of other aspects
of operating such a network, including address
The Mobile WiMAX network consists of the allocation and authentication. Figure 10 shows
access services network (ASN) and connectiv- an overview of the architecture.
ity services network (CSN). The core elements
in the ASN are the base station (BS) and ASN • SS/MS: Subscriber Station/Mobile Station
gateway (ASN-GW), which are connected over • BS: Base station, part of the ASN
an IP cloud. The functionality across the ASN-GW • ASN-GW: The ASN Gateway, part of the
and BS is split and signaled via R6. The ASN-GW ASN
provides security anchoring, traffic accounting, • CSN: The Connectivity Service Network
and mobility anchoring (and proxy) for the mobile • HA: Home Agent, part of the CSN
station (MS). The Mobile IP home agent (HA) in • AAA: AAA Server, part of the CSN
the CSN is used as a global mobility anchor, and • NAP: A Network Access Provider
Figure 8. DSRC spectrum band
8
Figure 9. Mobile WiMAX network architecture
neXT-generaTion Vehicular suitable platforms that would provide the desired

neTwork archiTecTure telematics and Internet services for vehicles.
For instance, the Intel Corporation has an-
This section discusses the architecture of the nounced plans to bring PCs to cars in a project
Next-Generation Vehicle Network, which is under called Connected Car PC Technology, and has
development at the University of Detroit Mercy indicated it at many events, including the Con-
by several researchers in collaboration (Mahfoud sumer Electronics Show in Las Vegas in 1998
et al., 2008). (Machen, 2009), where Intel demonstrated a Ford
Expedition loaded with a Pentium MMX-based
in-Vehicle personal computer computer. This computer provides voice- activated
satellite-based navigation, Internet access, cellular
Automotive manufacturers are planning to provide phone calls, games and DVD movies.
new Internet and entertainment services inside ve- The IBM infotainment system Imaj provides an
hicles. These services need to monitor and collect example of on-board computers. Imaj comprises of
data collection from increasing the requirement for three 12-inch LCD screens built into the car seats
computing capabilities onboard vehicles (InCode facing three rear-seat passengers. These screens
Telecom Group, 2009; Microsoft drives, 2004; provide access to radio, audio player, video player,
Mosra et al., 2004). The largest computer hard- fax, phone, e-mail, calculator, satellite mapping,
ware and software companies are now developing local information, word processing, web brows-
9
Figure 10. The WiMAX overview of the architecture
ing, and calendar functions all controlled either The .NET Connected Car uses the Windows
by mouse or by speech using IBM via-Voice Automotive operating system, which provides
technology (Ibm.com). computing services such as voice-activated com-
Software platforms are being developed for mands, and video and audio support for entertain-
in-vehicle computers to provide voice-activated ment. The .NET Connected Car also includes
commands. For instance, Microsoft Corporation built-in support for popular wireless technologies,
has launched the .NET Connected Car (Microsoft including Wi-Fi, General Packet Radio Service
drives, 2004), which provides the next genera- (GPRS), Bluetooth and Code Division Multiple
tion of software for the melding of computing Access (CDMA) (Microsoft drives, 2004).
and communications in the vehicle (Microsoft.
com, 2009). Microsoft describes Connected Car system architecture
as follows: “Tomorrow’s driver can expect better
information through systems built on the Microsoft Figure 11 shows the system architecture, which
.NET Connected Car software, a new generation of consists of three subsystems. The first subsystem
technology that connects individual cars to other is the vehicle, which includes the various networks
information systems through the Internet. Your and the on-board PC. Next is the wireless com-
next car may tell you that you should change the munication network, which provides a telephone
oil sooner than expected. It may also direct you to service, as well as a WAN (Wide Area Network)
the closest dealer, help schedule an appointment, connection between the PC and the Internet. The
notify you of oil change specials, and keep you third subsystem is the Internet, which provides a
informed on up-to-date service alerts.” (Microsoft database server for vehicle information.
drives, 2004).
10
Figure 11. A block diagram of the next generation vehicle network
Vehicle Network and the On-board PC the CAN nodes have a specially-developed Boot-
loader core in the nonvolatile memory of the MCU,
This architecture considers that all the intelligent thus enabling access to the MCU via the CAN bus
sensors, actuators and other control units inside through the CAN port of the MCU to reprogram the
the vehicle are implemented by CAN-capable Flash memory when required. This new feature is
microcontrollers. In this case, the CAN protocol essential for facilitating and easing the debugging
is implemented for networking these devices and reprogramming of the CAN nodes, as well as
together. Networking all these devices together for enabling this task to be performed remotely
in one network would enable them to exchange over the Internet. The CAN nodes should also be
data among themselves, while also allowing any able to operate in Wakeup mode, the nodes would
device to transmit data to an external network, or consume the minimum amount of battery power
be accessed from it to provide features such as when the vehicle is stopped, while still being able
remote diagnostics, monitoring, data collection to be woken up remotely, if required, to perform
and remote firmware upgrade. a diagnosis task.
This architecture makes various hardware and To reduce the cost and size, and to minimize
software changes to the vehicle CAN network in technology obsolescence, an on-board computer
order to make this network Internet-enabled. The generally does not have very high storage and
remote firmware upgrade feature of the CAN nodes processing capacity (Encyclopedia.com, 2009).
requires the most changes. The discussed archi- Additionally, this computer mostly acts as a
tecture requires the CAN modules to have Flash client depending on the remote server on the
memory for code, and an on-module or on-chip Internet, and therefore does not require very high
(built into the micro-controller unit (MCU)) Flash performance.
programming voltage generator. The firmware of
11
The In-vehicle, or on-board, PC plays several sets and the wireless PCMCIA cards for laptops
roles in this architecture. It is the master control and pocket PCs (Gsmworld.com, 2009).
unit for the vehicle, the TCU (Telematics Control GPRS has a theoretical maximum speed of
Unit) (Ibm.com) and a gateway for the different 171.2 kilobits per second (kbps) when using all
networks inside the vehicle. The gateway role pro- eight timeslots at the same time. The most impor-
vides connectivity among the in-vehicle networks tant feature of the GPRS is immediacy, which a
themselves, and between each of them and the dialup connection is not required. Users can access
Internet as necessary. Therefore, the PC requires the Internet immediately, since they are always
an interface to every network that needs to be con- connected due to the packet-switching technology
nected to it. Additionally, the PC should also have adopted by GPRS on the existing circuit-switching
an interface to the Internet to provide connectivity GSM (Gsmworld.com, 2009).
(gateway) between the vehicle network and the
Internet, as well as to provide regular Internet The Internet
access for browsing and entertainment.
An Internet server is required to store uploaded
The Wireless Network data from the vehicles. Each vehicle has its own
account on such server. The manufacturer and the
The wireless network in this architecture exists to service center can access to the vehicle’s date in that
provide communication, phone calls and Internet server. The Internet is also adopted to access the
access inside the vehicle. In this subsystem, the computer on board the vehicle and, consequently,
on-board PC is connected to the Internet via a to access the vehicle network that is connected to
WAN (Wide Area Network) Connection. This and monitored by that computer. Accessing the
WAN is a wireless Internet connection provided on-board computer from the Internet requires the
by any appropriate wireless communications car- implementation of strict security measures that en-
rier for cellular phones and similar devices. Speed sure appropriate access for authorized parties only.
and availability are the main factors to determine Vehicle data can be protected by implementing
the suitability of that carrier. The GPRS (General frameworks for data protection based on privacy
Packet Radio Service) (Gsmworld.com, 2009), and security technologies (Duri et al., 2002).
which is a packet-switching, non-voice service,
can be implemented as the wireless Internet
connection. It allows information to be sent and reFerences
received across a mobile telephone network, and
provides actual packet radio access and time- Duri, S., Gruteser, M., Liu, X., Moskowitz, P.,
division multiple access (TDMA) to users. Perez, R., Singh, M., & Tang, J. M. (2002).
GPRS is currently adopted to gain wireless Framework for security and privacy in automotive
access to many Internet applications and services telematics. In Mobile Computing and Networking,
wirelessly such as chat, textual and visual infor- (pp. 25-32).
mation including the download of audio, still and Eichler, S. (2007). Performance evaluation of the
moving images; web browsing; email; vehicle IEEE 802.11p WAVE communication standard.
positioning with collaboration from GPS; remote In Proceedings of IEEE Vehicular Technology
LAN access; and file transfer. These applica- Conference, (pp. 2199-2203).
tions can be accessed by several GPRS devices
(terminals), such as web-enabled cellular phone
12
Encyclopedia.com. (2009). Telematics is not a Kim, J. H., Seo, S. H., & Moon, T. Y. (2007). A
question of if, but when. Retrieved from http:// method of improving the reliability of the gateway
www.encyclopedia.com/doc/1G1-94875067. system by using OSEK/VDX. In Proceedings of
html International Conference on Control, Automation
and System, (pp. 2328-2833).
FlexRay Communications System Protocol Speci-
fication v2.1 Revision A. (2005). Machen. L. (2009). Technical marketing engineer,
“Intel drives in-vehicle solutions” handheld
FlexRay Consortium. (2009). Retrieved from
components division. Retrieved from http://
http://www.flexray.com
www.intel.com/technology/magazine/computing/
Gsmworld.com. (2009). Retrieved from http:// it04001.pdf
www.gsmworld.com/technology/gprs/index.
Mahfoud, M., Al-Holou, N., & Baroody, R. (2008).
shtml
Next generation vehicle network: Web enable. In
Ibm.com. (n.d.). Focus on the road ahead: IBM Proceedings of 3rd International Conference on
puts practical telematics within reach. Retrieved Information and Communication Technologies:
from http://www-306.ibm.com/software/perva- From Theory to Applications, (pp. 1-7).
sive/info/Telematics_within_reach_050404.pdf
Microsoft drives to Las Vegas. (2004). Connected
IEEE802.16-2005. (2006). Part 16: Air interface Concept Cars. Retrieved from http://www.win-
for fixed and mobile broadband wireless access dowsfordevices.com/news/NS5155857065.html
systems amendment 2: Physical and medium ac-
Microsoft.com. (2009). Windows Automotive.
cess control layers for combined fixed and mobile
Retrieved from http://www.microsoft.com/win-
operation in licensed bands and corrigendum 1.
dowsautomotive/wa5/default.mspx.
IEEE P802.11p/D3.0. (2007). Draft amendment
Mosra, S. R., Shanker, S., & Mahmud, S. M.
for wireless access in vehicular environments
(2004). An intelligent architecture for issuing
(WAVE).
intersection collision warnings (pp. 176-183).
InCode Telecom Group Inc. (2009). Telematics: National Defense Industries Association (NDIA).
How economic and technological forces will OSEK VEX Portal, (n.d.). Retrieved from http://
shape the industry in the U.S. Retrieved from www.osek-vdx.org
http://www.incodewireless.com/pdfMailer/files/
Pazul. K. (1999). An introduction to the CAN
Telematics_Position_Paper_v11.pdf
protocol that discusses the basics and key features.
Jiang, D., & Delgrossi, L. (2008). IEEE 802.11p: Microchip Application Note #AN713.
Towards an international standard for wireless ac-
Robert Bosch Gmb, H. Stuttgart, Germany, (1991).
cess in vehicular environments. In IEEE Vehicular
CAN Specification ver. 2.0.
Technology Conference, (pp. 2036-2040).
Sun, Y., Wang, F. Y., Wang, Z. X., Qiao, X., &
Johansson, K. H., Torngren, M., & Nielsen. L.
Wang, K. F. (2006). A scheduling algorithm for
(2003). Vehicle applications of controller area
vehicular application specific embedded operating
network. Technical Report Department of Signals,
systems. In Proceedings of IEEE International
Sensors and Systems, Royal Institute of Technol-
Conference on Systems, Man and Cybernetics,
ogy, Stockholm, Sweden, Department of Electrical
(pp. 2535-2540).
Engineering, Linkoping University, Sweden.
13
Wu, H., Fujimoto, R., Hunter, M., & Guensler, Xu, Y. N., Kim, Y. E., Cho, K. J., Chung, J. G., &
R. (2005). An architecture study of infrastructure- Lim, M. S. (2008). Implementation of FLexRay
based vehicular networks. In ACM MSWiM, communication controller protocol with applica-
Montreal, Canada, (pp.36-39). tion to a robot system. In Proceedings of 15ht
International Conference on Electronics, Circuits
Xiang, W., Richardson, P., & Guo, J. (2006). In-
and Systems, (pp. 994-997).
troduction and preliminary experimental results
of wireless access for vehicular environments
(WAVE) systems. In Proceedings of 3rd Annual
International Conference on Mobile and Ubiqui-
tous Systems, (pp.1-8).
14
15
Chapter 2
Network Applications
Yao-Chung Chang
National Taitung University, Taiwan, R.O.C.
absTracT
Information and Communication Technology (ICT) is concerned with all the technologies that manage,
process, and communicate information. It is also named as telematics, combining two words: telecom-
munications and informatics, which is widely used in the application of Global Positioning System
technology integrated with computers and in the mobile communications technology for automotive
navigation systems. Table 2.1 and Table 2.2 respectively list the telemetric applications from user’s
point of view and the practical applications of vehicular telematics. Four applications of the vehicular
network are discussed in this chapter. The first section introduces the vehicular network application
services. The second section discusses the vehicular network application management. The third section
provides the platform technologies of vehicular network application. Finally, future vehicular network
application and deployments are presented in the fourth section.
inTroducTion the wireless optimized TCP, the wireless optimized

HTTP, the SMS gateway, the push module and the
Service bundle provision systems (Choi et al., 2005; framework, as illustrated in Figure 1. Such a system
Han et al., 2005) enable users to search, download has three main components: gateway, framework
and install service applications that can be operated and world telematics protocol.
on a user terminal in an open telematics environment
(shown in Figure 1). The system has incorporated the • Gateway: The gateway allows developers
telematics gateway in the service providers including to write a telematics server application that
DOI: 10.4018/978-1-60566-840-6.ch002
Introduction of Vehicular Network Applications
Figure 1. Open telematics service
can be operated irrespective of the gateway The telematics portal framework (shown in
of the mobile network. Figure 2) is mainly divided into three components.
• Framework: The developers can write The first component is a provisioning service part
applications, without knowing about the that provides a telematics terminal with telemat-
details of integrating the related servers ics service bundles. The second component is a
distributed in networks, by utilizing APIs service bundle management module that provides
supported by the framework. functionality for deploying and managing service
• World Telematics Protocol (WTP) bundles. The third component is a repository mod-
(WTP1.0 Specification, 2004): WTP de- ule that not only supports other two components
fines a protocol to exchange messages be- and stores service bundles and related data, but
tween a telematics terminal in a vehicle and a also allows search and manage service bundles.
telematics service center. Telematics service Additionally, it supports various software provid-
developers and service providers can devel- ing protocols including JNLP, MIDP and J2EE
op and provide telematics services that do client provisioning.
not depend on devices and service carriers.
See Tables 1 and 2. Table 2. Practical applications of vehicle tele-
matics
Wireless vehicle safety

Vehicle tracking communications
Table 1. Telematics applications from user’s point Trailer tracking Emergency warning system
of view for vehicles
Cold store freight logistics Intelligent vehicle technolo-
Audio (CD / Radio) Telephone gies
Telephone Navigation Fleet management Car clubs
Video Speech recognition Satellite navigation Auto insurance
Short messaging (SMS) User interface for body Mobile data and mobile
electronic television
16
Proactive services (Gura et al., 2001) describe • User agents provide a user-defined appear-
a mobile application for traffic telematics based ance of user-level services. Thus, a single
on the JINI middleware. The main contribution look-and-feel interface encompasses the
proactively is the definition of a user agent that various interfaces of various services as
closely interacts with a service discovery and much as possible.
lookup facility. The lookup service is available in
each vehicle and provides a set of both user-level The Road-Look-Ahead (RLA) Service (Gura
services and lower-level services for accumulating, et al. 2001) is composed of all relevant elements
improving and distributing information obtaind of a proactive service listed in the above three
from various embedded sensors. The low-level statements. This service dynamically links a car,
services form the basis for context creation. To e.g. a truck as shown in Figure 3, advertising an
exploit these services proactively, a user agent is image from the local telematics system. To select
designed to perform the following tasks: an image that may be offered by many cars in
reach, the user agent needs to perform a selection
• User agents monitor available services on according to the context information, including
other vehicles. They are parameterized by the speed of the car and the direction, thus filter-
user profiles, and are selected based on ing out cars that are far away or come from the
user settings. opposite direction. The system architecture is
• User agents assess the lower-level floating based on Java (Sun Microsystems Inc., 2000) and
car data to generate the context based on JINI (Sun Microsystems Inc., 1999) technologies.
which decisions involve services. The concept of a service is essential to the JINI
system and gives the system various advantages.
Figure 2. The portal framework overview
17
First, the low-level context information such as Vehicular neTwork

position, direction and speed can be abstracted applicaTion managemenT
and encapsulated in a JINI location service that
abstracts from all specific details of the navigation Communications resources management (Pun-
devices. Second, JINI provides several mecha- noose et al., 2001) for advanced telematics ap-
nisms linked to mobility. plications is divided into three parts: external
Growing consumer demand for access informa- communications, internal communications and
tion in remote, mobile environments has sparked interface to resource manager. The architecture
interests in In-Vehicle Telematics Systems (IVIS). which supports mechanisms for internal and ex-
This work introduces the EmergeITS project ternal communications is shown in Figure 5.
(Reilly et al., 2002), which focuses on using IVTS Communications between the in-vehicle
in emergency fire services. In particular, this work components and the exterior pass through a
describes a distributed service-based architecture, router gateway. This routing module selects the
based on the JINI middleware technology, which appropriate external communication method for
can be employed to provide fault-tolerant applica- data connections according to the parameters
tion services to remote in-vehicle computers and such as cost, availability and QoS requirements of
mobile devices including Palm devices and WAP the applications. The resource manager interacts
phones. EmergeITS adopts the JINI middleware with this routing module to advise it of the QoS
(Figure 4) to provide a service-based architecture requirements of applications.
that can manage, configure and provide applica- Inter-agent communication within the vehicle
tion services to remote in-vehicle computers and must be flexible in order to facilitate the addition
Palm devices. of new software agents. A model that allows the
Figure 3. Road-look-ahead service
18
Figure 4. SunLabs’ concept of car software architecture
Figure 5. Management of internal and external communications
transfer of data without strict pre-definition (Raj- underlying infrastructure, which merely transports
kumar et al., 1995) is preferred, so that all agents it from the publishers to the subscribers. Publishers
that tap into this infrastructure can be publishers or post information on logical channels. Other agents
subscribers of information, or both. The informa- can be subscribers to this information.
tion that is being published is transparent to the
19
The resource manager can be a single or dis- ers) by decreasing maintenance and increasing
tributed entity. Client applications communicate convenience.
with the resource manager via network packets. The proposed SVMS adopts CDMA wireless
Therefore, the interface to the resource manager communication (Chan et al., 2003) to control and
can be embedded in the same processor as the manage a vehicle (Bjic et al., 2002). CDMA allows
client application or in a different machine within vehicle owners to access their vehicles anywhere
the vehicle. The only information required by the anytime. Additionally, a handset can be adopted
client is the location of the resource manager. as the remote-controller for the vehicle. VMP
A modern vehicle contains many electronic provides a graphic user interface that manages the
devices (Nolte et al, 2005), linked by Local Inter- vehicle and owner. The gateway plays in a relay
connect Network (LIN), Controller Area Network role. The developed gateway is an intermediate
(CAN) and FlexRay (FlexRay Communications among handsets, VMPs and vehicular devices,
System Protocol Specification v2.1 Revision since this system supports LIN, CAN, FlexRay
A, 2005), to fulfill the needs of customers and and CDMA interfaces. Thus, the proposed SVMS
to enhance the performance of the vehicle. The not only achieves telematics and a ubiquitous
smart vehicle management system (SVMS) (Seo environment within a vehicle, but also manages
et al., 2007) is composed of a gateway, handset electrical components.
and vehicle management program (VMP). The Figure 6 illustrates the VMP running on a PC
increasing number of electronic devices enables with a RS-232 interface. The CDMA modem is
automatic control and management of vehicles. linked to the PC via RS-232. The CDMA modem
In particular, automatic management of a vehicle transfers the received messages to the VMP, and
provides significant benefits to owners (driv- the commands sent to the gateway. The handset
Figure 6. Smart vehicle management system prototype design
20
accesses the gateway directly by using the SMS. 2009) and AMI-C Specification (AMI-C, 2002).
The gateway is linked with three embedded sys- Telematics technology, which adopts various tele-
tems. For instance, embedded system #3 operates communications such as DMB, CDMA, WRAN
several LEDs based on VMP commands or hand- and DSRC, is increasingly used to provide services
set messages. Each embedded system is connected for in-vehicle telematics systems or context aware
using CAN and LIN. Additionally, all operations automotive telematics (Zhang et al., 2004). This
of each process are printed in VMP and saved in service bundles provide in-vehicle applications
a file by a log writer. such as vehicle status monitoring service, vehicle
The application presentation of telematics is no tracking and local advertising, can be realized us-
restricted to the classical embedded control system, ing. This architecture allows in-vehicle terminals
but instead it covers a broad range from driver as- to provide various telematics services to improve
sistance to infortainment and vehicle information safety of drivers.
management. Management of various telematic
applications requires new application managers
(Kim et al., 2006) for the automotive domain to Vehicular neTwork
support the specific software components, cor- applicaTion plaTForm
responding to the vehicle status, and eventually
combine the system parts to form a single reliable Figure 8 shows the architecture and design of an
and manageable system. This work develops a OSEK/VDX (O’Donnell, 2003) (includes specifi-
novel architecture for application management cations for OSEK Operating System (OS) (OSEK
in in-vehicle software (shown in Figure 7). This Group, 2005), OSEK Communication (COM)
architecture is based on international standards, (OSEK Group, 2004), OSEK Network Manager
including OSGi (Open Service Gateway Initia- (NM) (OSEK Group, 2004) and OSEK Implemen-
tive, 2003) framework (two major platforms are tation Language (OIL) (Zheng et al., 2004)) and
Prosyst’s mBedded Server (Prosyst, 2009) and OSGi-based Embedded Software Platform (Sun et
Gatespace’s e-Services Platform (Gatespace, al., 2007). Three components are introduced: The
Figure 7. Architecture of application manager
21
remote vehicular service platform component has manufacturers or third party developers. Vehicular
high computation ability and huge data storage. correlation services mainly provide facilities such
The vehicle/home interactive platform component as GPS location and information transmission
is another OSGi-based framework gateway. The management. Transferring information accurately
vehicle driver can interact with it via the wireless in real-time is a key to the quality of service.
connection and manage home/office appliances Individual information services provide services
while driving. The roadside system component including entertainment search and download,
refers to ITS infrastructure such as intelligent information service and vehicle/home interac-
traffic light controllers and communication base tive service.
stations. Telematics open portal applications zone
Figure 9 shows OSGi-based remote vehicular (TOPAZ) (Lee et al. 2005) is under development
service platform (Wang et al. 2004). The vehicular by the IBM Ubiquitous Lab. Figure 10 describes
integration services mainly combine diagnostics two telematic applications based on TOPAZ plat-
and prognostics to form the optimal vehicular form (Choi et al., 2006). These applications are a
control algorithms download services. These call-taxi service and an emergency rescue service
services are generally provided by automotive for patient in u-health scenarios. It builds three
Figure 8. Architecture of the OSEK/VDX and OSGi-based embedded software platform
22
Figure 9. OSGi-based remote vehicular service platform
different applications end-to-end within a short rule-based engine for event detection in TOPAZ
time period and minimal manpower. Call taxi to determine the critical status of the patient.
application provides the nearest available taxi to The growth in Internet subscribers has ac-
a customer through vehicle tracking information celerated at an exponential pace. Wired Internet
until the assigned taxi arrives at the customer’s access serves subscribers only at attached points,
location. The main parts of the application service while wireless communications provide enjoy
are service clients, call center server and vehicle ubiquitous/pervasive Internet access. Recent
dispatch server (VDS). The call center server does advances in wireless inter-vehicle communica-
not directly connect to TOPAZ servers, but instead tion systems have generated major opportunities
interacts with VDS and a customer to transfer the for the deployment of a wide variety of services
requested data between them. to vehicles (Dikajakos et al., 2008). Telematics
Emergency rescue service for patient is a pre- provide innovation and new technology in the
ventive service that monitors biological status of automotive sector, while significantly improving
patient using ubiquitous sensors attached to the the driving experience (Grymek et al., 2007).
patient body, predicts the critical situation of the This investigation combines wireless services
patient, and dispatches the emergency vehicle to with human-computer interaction to provide a
the patient’s location. The Health management ubiquitous/pervasive computing environment for
server (HMS) was developed to interact with telematics services.
patients’ devices using the same VDS adopted IETF defines mobile IP facilities to provide
in the call taxi application, but to dispatch emer- roaming methods in wireless computing. The
gency vehicles rather than taxis. The proposed NEMO (NEtwork MObility) (Emst, 2006; De-
emergency rescue service application adopts varapalli et al., 2005) working group has improved
the spatiotemporal event detection environment roaming schemes to enable subscribers to use mo-
STEDE (Munson et al., 2005; Lee et al., 2005) bile routers for data transmission without worrying
about service environments (Tseng et al., 2007).
23
Figure 10. Application deployment in TOPAZ
A mobile router can be installed on any vehicle. receiving, sending and processing, is established
The vehicular router communicates with external in the vehicular router. This work also enhances
vehicular networks. Telematics application users mobile IPv6 technology and network mobility.
within a moving vehicle do not need to perform
handoffs individually. The entire vehicle is a sub-
network and performs handoff procedures using FuTure Vehicular neTwork
the installed vehicular router. Figure 11 illustrates applicaTion and deploymenT
the Mobile telematics computing environment.
This study develops an embedded WiMAX- A client-server application and Internet based ap-
based network mobility system on an Intel IXP425 plication were developed. Users performed various
network processor platform to deploy vehicular remote experiments on a mobile robot, focusing
routers for vehicular network telematics comput- on motor control, obstacle avoidance and image
ing (Chen et al., 2009). A mobile router can be processing, and applying these features to trajec-
installed on any vehicle. The vehicular router tory control using the Common Gateway Interface
is responsible for communicating with external (CGI) to access the robots (Backes et al., 2000;
vehicular networks. When a vehicle moves, Schilling, 2001). The remote user is connected
telematics users within the vehicle do not need via the Internet to a dedicated computer, which
to perform handoffs individually. The whole controls and monitors the mobile robot (Popescu
vehicle is a sub-network and performs handoff et al. 2008). The client was programmed in Java,
procedures using the installed vehicular router. while server was programmed in Visual Basic 6.
The concept of mobile telematics computing The Visual Basic server application handles the
environment is shown in Figure 11. The mobile application administration module, which involves
IPv6 scheme, which includes packet formatting, user management, and communication with the
24
Figure 11. Mobile telematics computing
control board and microcontroller program editor. prediction with floating car data is an example of
An AXIS 205 network video camera sends live emerging telematic applications that exploits of
video of the mobile robot to the user’s browser. global networking capabilities (Chen et al., 2004).
The Internet provides a major opportunity for A middlebox is a device performing a service that
robotics education by enabling the development needs application logic, but is executed in the
of remote laboratories, where students can gain network (shown in Figure 13). The aggregation
practical experience deploying hardware resources function receives UDP packets, processes them
efficiently. The system reduces the time and space as described above, and generates asynchronous
constraints normally associated with the traditional XML/RPC messages. The function then sends
laboratory, thus making equipment available to these messages to the server. XML/RPC is an
more students. RPC protocol that utilizes HTTP as the transport
The InternetCAR project (Ernst et al., 2003; protocol and XML for data representation. The
InternetCAR Project, 2003) develops the archi- front-end server was implemented as a Servlet
tecture needed to connect a car to the Internet, but running within the WebSphere Servlet engine.
does not address the problems of infrastructure A Mobile Agent-based Middleware for ve-
to enable the services. Telematics functions were hicle telematics is proposed in (Guo et al., 2007).
configured on a dedicated device, commonly The JNomad, a framework that integrates JINI
called a middlebox (Srisuresh et al., 2002). These technology with mobile agents, is developed
middleboxes adopt programmable network pro- to meet these deployment challenges. JINI is a
cessors in order to attain the required processing novel technology that considers the pervasive,
and forwarding speeds, while communicating ubiquitous, and dynamic distributed computing
with each other and the back-end server using requirements within a single architecture. To ad-
standard middleware components. Urban traffic dress the mobility and frequent network partition
25
Figure 12. Telematics architecture for mobile robot
Figure 13. Overview of middleboxes
problem, this investigation designed and imple- sequence of interactions among different compo-
mented a Code-on-Demand Service Model using nents within the framework is shown in Figure
JINI technology, and developed a Mobile-Agent 14. Using the JINI architecture, virtually any
based Service Model. service may interact freely in a network without
JNomad is a mobile agent-based framework the requirement for complex protocols, messaging
consisting of three major components. The first drivers, operating systems or cabling.
component is composed of mobile agents, i.e., Security issues are also very important topics
entities that are expected to perform various tasks. in vehicular applications. A general architecture
The second component contains the mobile agent for a trusted context provider is proposed in
hosts, which provide the platform executed by the (MOSQUITO Project, 2005). Different security
mobile agents. The third component comprises mechanisms can be used in different situations
lookup and location services that provide dynamic in this architecture. Information Connector is
service discovery and location updates. A typical presented for vehicular environments. A general
26
Figure 14. The agent-based framework using JINI technologies
Figure 15. Vehicular security architecture
27
architecture for security in vehicular environment devices are likely to access services using multiple
is presented in (Nowey et al., 2006). Figure 15 methods and technologies.
illustrates the overall architecture of vehicular se-
curity architecture (Nowey et al. 2007), including
Service Plane, Security Middleware and Context reFerences
Aware Applications.
AMI-C. (2002). Software API Specifications-
• Service plane: The security layer is CORE APIs. retrieved September 20 2009, from
configured using the security manager, http://www.ami-c.org
which provides the appropriate interface. Backes, P., Tso, K., Norris, J., Tharp, G., Slostad,
Additionally, the security manager sets the J., Bonitz, R., & All, K. (2000). Internet based
address of the position based routing en- Operations for the Mars Polar Lander Mission.
tity, thus protecting the location privacy of In The 2000 IEEE International Conference on
the users. Robotics & Automation, (pp. 2025-2032).
• Security middleware: The three major
parts of the security middleware are trust Bjic, E., & Chaxel, F. (2002). Auto-ID Mobile
establishment services, revocation services Information System for Vehicle Life Cycle Data
and privacy services. Management. IEEE Systems . Man and Cyber-
• Context aware applications: Context- netics, 4, 6.
aware applications perform actions such as
Chan, Alex Y. M., & Lu, W. P. (2003). Architecture
issuing a warning to the driver or peer appli-
for Wireless Access in Vehicles. IEEE Vehicular
cations based on the context information.
Technology Conference, 5, 3336-3340.
BMW, in collaboration with Telematics Service Chen, A., Jain, N., Perinola, A., Pietraszek, T.,
Providers (TSPs) Connexis (Connexis, 2009) and Rooney, S., & Scotton, P. (2004). Scaling Real-
WirelessCar (WirelessCar, 2009), has developed Time Telematics Applications using Program-
a new telematics framework and a technology- mable Middleboxes: A Case Study in Traffic Pre-
neutral telematics protocol to bring new flexibility diction. In First IEEE Consumer Communications
and scalability to the automotive sector called Next and Networking Conference, (pp. 388-393).
Generation Telematics Protocol (Next Generation
Chen, J. L., Chang, Y. C., Lin, Y. S., & Du, H.
Telematics Protocol, 2009). Vehicle manufacturers
W. (2009). Embedded WiMAX-based Vehicular
have historically provided proprietary services to
Router for Telematics Computing. In The 11th
customers, and have been dependent on a single
International Conference on Advanced Commu-
TSP for delivery of these services in a specific
nication Technology, (pp. 1857-1862).
market. This supply chain inflexibility has created
difficulties for providers in gaining economies of Choi, D. H., Ro, S. H., Lee, K. S., Lee, W. N.,
scale and promoting their products and services. Choi, S. Y., & Pae, Y. W. (2006). Telematics Ap-
Although an open and standardized approach plications Based On TOPAZ Platform. In IEEE
to delivering services has clear benefits for the 6th International Conference on Computer and
marketplace, previous standardization efforts Information Technology, (pp.262-268).
have encountered barriers to adoption, because
they have focused on replacing existing protocols
instead of integrating them. The proliferation of
new technologies suggests that future in-vehicle
28
Choi, J. W., Han, W. Y., Kim, C. S., & Kwon, O. Guo, J. H., & Xing, G. M. (2007). Using Mobile
C. (2005). Open Telematics Services Deployment Agent-based Middleware to Support Distributed
on the Gateway and Framework Independent on Coordination for Vehicle Telematics. In 21st In-
Mobile Networks. In International Conference ternational Conference on Advanced Information
on Wireless Networks, (pp.374-379). Networking and Applications Workshops, (Vol.
2, pp.374-379).
Connexis website. (n.d.). Retrieved September
20, 2009, from http://www.connexis.com/ Gura, N., Held, A., & Kaiser, J. (2001). Proactive
Services in a Distributed Traffic Telematics Ap-
Devarapalli, V., Wakikawa, R., Petrescu, A., Thu-
plication. Workshop on Mobile communication
bert, P. (2005). Network Mobility (NEMO) Basic
over Wireless LAN: Research and Applications.
Support Protocol RFC 3963.
Retrieved from http://www.informatik.uni-ulm.
Dikajakos, M. D., Florides, A., Nadeem, T., & de/rs/projekte/core/
Iftode, L. (2008). Location-Aware Services over
Han, W. Y., Kwon, O. C., Park, J. H., & Kang,
Vehicular Ad-Hoc Networks using Car-to-Car
J. H. (2005). A Gateway and Framework for In-
Communications. IEEE Journal on Selected
teroperable Telematics Systems Independent on
Areas in Communications, 25(8), 1590–1602.
Mobile Networks. ETRI Journal, 27(1), 106–109.
doi:10.1109/JSAC.2007.071008
doi:10.4218/etrij.05.0204.0038
Emst, T. (2006). Network Mobility Support Goals
Internet, C. A. R. Project (2003). InternetCAR
and Requirements.
webpage at WIDE. Retrieved September 20, 2009
Ernst, T., Uehara, K., & Mitsuya, K. (2003). Net- from http://www.sfc.wide.ad.jp/InternetCAR
work Mobility from the InternetCAR Perspective.
Kim, M. J., Choi, Y. J., Moon, Y. B., Kim, S. J., &
In 17th International Conference on Advanced
Kwon, O. C. (2006). Design and Implementation of
Information Networking and Applications, (pp.
Status based Application Manager for Telematics.
19–26).
IEEE 8th International Conference on Advanced
FlexRay Communications System Protocol Communication Technology, (pp. 20-22).
Specification v2.1 Revision A, (2005). Gatespace
Lee, S. W., Munson, J., Lee, D. R., Thompson,
website. (n.d.). Retrieved September 20 2009, from
G., & Park, J. (2005). A Rule-based System for
http://www.gatespacetelematics.com/
Sense-and-Respond Telematics Applications.
Gerlach, M., & Fokus, F. (2007). Trust for Ve- In The First Korea/Japan Joint Workshop on
hicular Applications. In 8th International Sym- Ubiquitous Computing & Networking Systems,
posium on Autonomous Decentralized Systems, (pp 109-114).
(pp. 295-304).
Lee, W., Tak, Y., Byun, W. H., & Pae, Y. W. (2005).
Grymek, L., Singh, S., & Pattipati, K. (2007). Toward an Open Telematics Infrastructure. IPSJ
Vehicular Dependence Adds to Telematics’ Al- SIG Technical Reports, (pp. 443-448).
lure. IEEE Potentials, 26(2), 12–16. doi:10.1109/
MOSQUITO Project. (2005). Mosquito Architec-
MP.2007.343053
ture. White Paper. Retrieved September 20, 2009
from http://www.mosquito-online.org/
29
Munson, J., Lee, S. W., Lee, D. R., Wood, D., Punnoose, J., Tseng, S., Wang, S., Nikitin, V.,
Thompson, G., & Cole, A. (2005). A Rule-based & Schlesinger, T. E. (2001). Communications
System for Sense-and-Respond Telematics Resources Management for Advanced Telematics
Services. Workshop on End-to-end, Sense-and- Applications. Transportation Society Confer-
Respond Systems, Applications and Services, ence.
(pp. 31-36).
Rajkumar, R., Gagliardi, M., & Sha, L. (1995).
Next Generation Telematics Protocol website. The Real-Time Publisher/Subscriber Inter-Process
(n.d.). Retrieved September 20 2009, from http:// Communication Model for Distributed Real-Time
www.ngtp.org Systems: Design and Implementation. Real-Time
Technology and Applications Symposium, (pp.
Nolte, T., Hansson, H., & Bello, L. L. (2005).
66–75).
Automotive Communication – Past, Current and
Future. In 10th IEEE Conference on Emerging Reilly, D., & Taleb-Bendiab, A. (2002). A Service-
Technologies and Factory Automation, 1, 985- Based Architecture for In-Vehicle Telematics
992. Systems. In the 22nd International Conference
on Distributed Computing Systems Workshops,
Nowey, K. P. T., & Mletzko, C. (2006). Towards
(pp.741-742).
a Security Architecture for Vehicular ad hoc Net-
works. In First International Conference on Avail- Schilling, K. (2001). Virtual Laboratories for
ability, Reliability and Security, (pp. 374-381). Mobile Robots. IFAC Workshop on Internet Based
Control Education, (pp. 115-120).
O’Donnell, M. (2003). OSEK/VDX: Driving
the Standard for Optimized Embedded Systems. Seo, S. H., Moon, T. Y., & Kim, J. H. (2007). Smart
Embedded Computing Design. Vehicle Management System by using Gateway,
Hand-set and VMP. International Conference on
Open Service Gateway Initiative. (2003). OSGi
Control, Automation and Systems, (pp.17-20).
Specification Release 3. Retrieved September 20
2009, from http://www.osgi.org Srisuresh, P., Kuthan, J., Rosenberg, J., Molitor,
A., & Rayhan, A. (2002). Middlebox Communica-
OSEK Group (2004). OSEK/VDX Network Man-
tion Architecture and Framework. IETF Internet
agement Concept and Application Programming
Draft, retrieved September 20 2009 from http://
Interface, Version 2.5.3.
www.jdrosen.net/papers/rfc3303.txt
OSEK Group (2004). OSEK Implementation
Sun, Y., Huang, W. L., Tang, S. M., Qiao, X., &
Language, Version 2.5.
Wang, F. Y. (2007). Design of an OSEK/VDX
OSEK Group (2005). OSEK/VDX Operation and OSGi-based Embedded Software Platform
System, Version 2.2.3. for Vehicular Applications. IEEE International
Conference on Vehicular Electronics and Safety,
Popescu, D., & Selisteanu, D. (2008). Web based
(pp. 1-6).
Telematics Application for Robotics. In the 3rd In-
ternational Multi-Conference on Computing in the Sun Microsystems Inc. (1999, November). JINI
Global Information Technology, (pp. 19-24). Specification v1.0.1. Retrieved from http//:www.
sun.com/JINI/
Prosyst website. (n.d.). Retrieved September 20
2009, from http://www.prosyst.com/
30
Sun Microsystems Inc. (2000). Java Media Frame- WirelessCar website (n.d.). Retrieved September
work API Guide, 2000. Retrieved from http://java. 20 2009, from http://www.wirelesscar.com/
sun.com/products/java-media/jmf/
WTP1. 0 Specification (2004). Electronics and
Tseng, Y. C., Chen, J. J., & Cheng, Y. L. (2007). Telecommunications Research Institute (ETRI),
Design and Implementation of a SIP-based Mo- Korea.
bile and Vehicular Wireless Network with PUSH
Zhang, D., Wang, X. H., & Hackbarth, K. (2004).
Mechanism. IEEE Transactions on Vehicular
OSGi based Service Infrastructure for Context
Technology, 56(6), 3408–3420. doi:10.1109/
Aware Automotive Telematics. IEEE 59th Vehicu-
TVT.2007.906986
lar Technology Conference, 5, 2957-2961.
Wang, F. Y., Li, J. X., & Liu, X. J. (2004). OSEK/
Zheng, N. N., Tang, S., Cheng, H., Li, Q., Lai, G.,
VDX and OSGi-based Embedded Platform for Ve-
& Wang, F. W. (2004). Toward Intelligent Driver-
hicular Software and its Key Techniques. Chinese
Assistance and Safety Warning Systems. IEEE
High Technology Letters, (pp.131-136).
Intelligent Systems, 19(2), 8–11. doi:10.1109/
MIS.2004.1274904
31
32
Chapter 3
Introduction to ITS and NTCIP
Da-Jie Lin
Feng Chia University, Taiwan, R.O.C.
Chyi-Ren Dow
absTracT
Intelligent Transportation Systems (ITS) combines high technology and improvements in information sys-
tems, communication, sensors, and relevant mathematical methods with the conventional world of surface
transportation infrastructure to increase the capacity of transportation systems and to improve the level
of services. There are four major goals of ITS, including safety, environmental protection, efficiency, and
economy. NTCIP (NTCIP Standard 9001, 2002; DISA et al., 1997) is a set of communications protocols
and data definition standards designed for various needs of ITS services and applications. The key goals
of the NTCIP open-standards effort are interoperability and interchangeability. Interoperability refers
to the ability for multiple devices to work together as a single system and interchangeability refers to
the ability to use multiple brands of a device on the same communications channel. Accompanying the
social and economic development, traffic congestion and delay have become major issues in most areas
around the world. How to use readily available technologies to increase the capacity of transportation
systems and to improve the level of service has become one of major solutions to solve transportation
problems that people are facing. This is the motivation of Intelligent Transportation Systems develop-
ment. NTCIP is a set of communications protocols and data definition standards designed for various
needs of ITS services and applications. These standards are intended to handle these needs in the two
areas: Center-to-Field (C2F) and Center-to-Center (C2C) communications.
DOI: 10.4018/978-1-60566-840-6.ch003
inTroducTion of advanced communication, sensing, and infor-

mation processing technologies. When integrated
Intelligent Transportation Systems (ITS) (ITS into the transportation system’s infrastructure,
Standard et al., 2006; ITE et al., 2003; RITA et and into vehicles themselves, these technologies
al., 2002) combines high technology and improve- relieve congestion, improve safety, and enhance
ments in information systems, communication, productivity.
sensors, and relevant mathematical methods with Overall, the goal of ITS is to utilize advanced
the conventional world of surface transportation technologies to increase the efficiency of limited
infrastructure. transportation resources, to increase user conve-
nience and the living quality. The content of ITS
definition of intelligent includes technologies and policies. The integration
Transportation systems of both technologies and policies and the applica-
tion of the integration are the core of ITS.
Intelligent Transportation Systems are defined There are four major goals of ITS: safety, en-
commonly as follow: systems utilizing newly vironmental protection, efficiency, and economy.
developed information and communications The details of these major goals are listed in
technology to transportation infrastructure, man- Table 1.
agement systems, and vehicles to improve safety, A dedicated website, ITS Benefits, Costs
efficiency, travel time and comforts and reduce and Lessons Learned Databases (http://www.
vehicle wear, delay, and fuel consumption. benefitcost.its.dot.gov/) provides updated and
Intelligent transportation systems include sev- comprehensive information about the benefit and
eral systems using different technologies, such as cost of ITS. The benefit of ITS can be classified
traffic signal control systems; vehicle allocation in two application approaches: (1) Intelligent
and navigation systems; container management
systems; changeable message signs (CMS); au-
tomatic number plate recognition (ANPR) or law Table 1. Major goals of ITS
enforcement equipments (speed cameras, CCTV, Goals Objectives
surveillance cameras, ... etc.) to more advanced 1. Safety 1. reduced incident frequency
applications such as traffic control centers that
2. reduced incident damages
integrate live data and feedback from a number
2. Environmental 1. less air pollution
of sources, such as microwave/infrared vehicle Protection
2. less greenhouse effects
detectors, parking guidance and information
3. less noise
systems; weather information sensors, analyze
4. less fuel consumption
all information through different models and
5.less new facilities & construc-
then control the signal systems in the facilities tions
to improve the traffic condition. 3. Efficiency 1. less travel time
2. increased capacity
goals and benefits of intelligent 3. reduced operating costs
Transportation systems
4. higher user satisfaction
4. Economy 1. increased industry production
Intelligent transportation systems provide a set of
2. more relevant job opportuni-
strategies for the transportation problems related ties
to safety and congestion, while accommodating (Source:Taiwan logistics management yearbook -Intelligent
travel demands of users and freight through the use Transportation Systems, 2001)
33
Infrastructure and (2) Intelligent Vehicles. In Advanced Traffic Management

terms of infrastructure part, the benefit of ITS Systems (ATMS)
includes better control in arterial and freeway
management, vehicle crash prevention and safety, Advanced Traffic Management Systems is the
road weather management and warning systems, core and foundation of Intelligent Transporta-
transit systems in improved level of service, traf- tion Systems. ATMS uses technologies such as
fic incident management to reduce any incident detecting, sensing, communications, and control
impacts, better emergency management, labor- and to collect the roadside data that will be transmitted
error-free electronic payment and precise pricing, back to a traffic control center. Data from different
readily available traveler information to reduce sources will be analyzed and traffic control center
user costs, ...etc. will determine their control strategies based on
On the other hand, intelligent vehicles increase the result of analysis in order to systematically
driving safety through new technologies such as optimize the whole network. The analysis results
collision avoidance and driver assistances such will be used to control traffic (kight or signau) and
as car navigation systems and driver monitoring the analysis results will be relayed to the public.
systems. ATMS increases the efficiency of transportation
networks as well.
subsystems of intelligent ATMS emphasizes the integration with
Transportation systems other systems and real-time traffic data. Related
technologies and applications includes traffic
There are commonly discussed nine ITS subsys- signal control, ramp control, automatic incident
tems that have their own functions and required detections, automatic vehicle locations (AVL),
technologies. They work either jointly or inde- changeable message signs (CMS), Geographic
pendently, depending on the scenarios. Information Systems (GIS), Weigh-in-Motion
(WIM), Automatic Vehicle Classification (AVC),
Electronic Toll Collection (ETC), Automatic
Vehicle Identification (AVI), …etc.
Figure 1. Advanced traffic management systems (Source:ATMS)
34
As shown in Figure 1 illustrates the interac- of their time to increase their productivity (see
tions among control center, infrastructure, and Figures 2 and 3).
vehicles. Relevant applications and technologies include
Highway Advisory Radio (HAR), Global Position-
Advanced Traveler Information ing Systems (GPS) and Geographic Information
Systems (ATIS) Systems (GIS), On Board Units (OBU), Wireless
communications, and Integrated Service Digital
Advanced Traveler Information Systems are Network (ISDN), etc..
designed to provide users the information they
need to plan their trips. The factor considered in Advanced Public Transportation
trip planning consists of transportation modes, Systems (APTS)
costs, schedules, accessibilities, connecting trans-
portation, etc.. ATIS utilizes the data collected Advanced Public Transportation Systems (APTS)
from different sources and processes these data use the technologies of ATMS, ATIS and AVCSS
to make them available and understandable to and apply them in the field of public transporta-
users. The platform the ATIS uses to deliver the tion to improve the service quality, efficiency, and
information to users includes TV at home, radio attractiveness of existing public transportation
in vehicles, changeable message sign (CMS) on services such as buses or trains. Users, instead
the roadside or in transit stations, internet (Feit of being delayed due to the uncertainty of the
et al., 1993) (wired or wireless connection), cel- transit systems, have better utilization of their
lular phones, etc.. Through this real-time, updated time. On the other hand, transit agencies have
travel information, the uncertainties in travel have a better control over their own fleet to improve
been greatly reduced and users have better control their management.
Figure 2. GPS navigation systems - to provide route guidance (Source: Japan ITS HANDBOOK 2006-
2007)
35
Figure 3. Changeable message sign (Source: http://www.flickr.com/photos/moogal/442796974/)
In the case of buses, dynamic bus information reasons of transportation incidents. Increasing the
systems, one APTS system for example, provide level of automation in driving will influentially
passengers detailed information such as route reduce the incident possibilities. For example, the
information, real-time bus location, estimated onboard driver monitoring systems will monitor
bus arrival time, connecting information (other the physical condition of the driver when he is
bus routes or modes) so passengers can plan their executing his driving duty. Another example is
travel in a more efficient manner. collision avoidance system that will work to slow
Related applications and technologies for down the vehicle when it is too close to the vehicle
APTS include Automatic Vehicle Monitoring in the front and the driver is not responding to this
(AVM), Automatic Vehicle Locating (AVL), situation accordingly (Figures 5 and 6).
wireless communications, Electronic Fare Pay- Related AVCSS technologies and applications
ment (EFP), bus scheduling systems, etc. See include automatic parking systems, collision
Figure 4. avoidance, communications between vehicle and
facility, driving behavior monitoring, human fac-
Advanced Vehicle Control and tor engineering, etc.
Safety Systems (AVCSS)
Commercial Vehicle Operations (CVO)
Advanced Vehicle Control and Safety Systems
integrate technologies such as sensors, onboard Commercial Vehicle Operations applies ATMS,
computers, communications, electronics, and ATIS, and AVCSS technologies in the commercial
control to improve the vehicles and transporta- vehicle operations such as trucks, towing trucks,
tion facilities in terms of safety, capacity, and taxies, paramedics, cranes, and other commercial
level of service. fleets. The onboard GPS provide the real-time
The advanced vehicle control and safety tech- vehicle location information and wireless com-
nologies have been deployed in vehicles to assist munication transmit the information back to
drivers to improve their driving skills, to reduce control centers or dispatching centers. This loca-
the likelihood of human errors and to increase the tion information is an important input to a fleet
safety, especially fatigue has been one of the major management system. After processing vehicle
36
Figure 4. Dynamic bus information systems
Figure 5. Onboard monitoring devices (Source: The Intelligent Transportation Systems of future)
Figure 6. Driver monitoring systems (Source: The Intelligent Transportation Systems of future)
37
location information, the fleet management system to make itself a better solution such as vehicle
can assign their commercial vehicles in a more no need to slow down, no incurred delays, lower
efficient way to improve their productivities. long-term operating cost, no need to handle cash,
Logistics companies, for example, see CVO and increased safety and facility utilization. See
as a critical system: freight being handled will Figure 8.
be efficiently allocated in proper vehicles and be The other use of EPS is e-tickets for transit
precisely tracked once they are en route. CVO systems such as buses, taxies, parking, and airlines.
enables them to reduce their costs while providing If being combined with other payment systems
better delivery services (Figure 7). successfully, users only need to carry one card
Related CVO technologies and applications to travel around without carrying a lot of cash or
include Automatic Vehicle Monitoring (AVM), waiting in a manual operation line.
Automatic Vehicle Locating (AVL), Weight- Related EPS technologies include automatic
in-Motion (WIM), Electronic Toll Collection vehicle identification (AVI), microwave/infrared
(ETC), Automatic Vehicle Identification (AVI)
and Automatic Cargo Identification (ACI), fleet
dispatching systems, etc. Figure 8. EPS (ETC) in Taiwan
Electronic Payment Systems (EPS)
Electronic Payment Systems or Electronic Toll

Collection (ETC) charges drivers/vehicles through
the communications between onboard units
(OBU) and roadside units. Conventionally tolls
are collected by hands and that requires vehicles
slowing down and preparing exact amount of fares
otherwise incidents or delays will be incurred.
On the contrary, EPS has the following features
Figure 7. Commercial vehicle operations
38
sensing, Radio Frequency IDentification (RFID), route with the minimum delays, and purposely
electronic money, etc. See Figure 9. adjusts the traffic signals along the route. Other
agencies such as police will be notified as well
Emergency Management to assist the rescue actions. EMS also determines
Systems (EMS) the impact area and evacuates people within the
area if necessary (Figures 10 and 11).
Emergency Management Systems is a system Related EMS applications and technologies
designed for emergencies. If an emergency such consist of Automatic Vehicle Location (AVL),
as a fire happened, EMS determines the dispatch- Geographic Information Systems (GIS), auto-
ing and routing of fire trucks and ambulances, matic incident detection, Highway Advisory
and these vehicles will be given priority to go Radio (HAR), etc.
to the scene. EMS utilizes the real-time traffic
data from control centers, determines the best Vulnerable Individual Protection
Systems (VIPS)
Figure 9. Automatic vehicle identification (Source: As shown in Figure 12, Vulnerable Individual
CECI, Taiwan, 2007) Protection Systems (VIPS) protects the safety
of handicaps, senior people, pedestrians, bikers,
motorcyclists, etc. For example, dedicated traffic
signals are designed for the blinds and handheld
navigation devices on a bicycle or motorcycle
that give warnings when a big truck is close from
the behind.
Related VIPS technologies include navigation
systems, voice-warning traffic signals, etc.
Figure 10. Automatic image detection system to identify incidents
39
Figure 11. Automatic incident detection-tunnel fire (Source: CECI, Taiwan, 2007)
Figure 12. Vulnerable individual protection systems
Information Management and agencies and facilitate decision making. Users

Systems (IMS) use this information to plan their trips to increase
safety, efficiency, and reduce the travel time and/
As shown in Figure 13, Information Management or travel expenses. Agencies use this information
Systems is the backbone of the Intelligent Trans- to better manage their vehicles, terminals and to
portation Systems. IMS use all relevant informa- reduce the operating cost. Authorities use this
tion and communication technologies to collect, information to design, modify or fine-tune their
transmit, archive, process, analyze and share data. strategies, facilities to optimize the system per-
Data such as travel time, travel speed, vehicle formance. Without IMS, ITS designs, strategies,
volume, travel origin-destination, ridership, ve- approaches and subsystems cannot be realized to
hicle location, etc. are further analyzed to become the extent that they were designed to be.
usable information that will be delivered to users
40
Figure 13. Traffic control center and IMS interface (Source: CECI, Taiwan, 2007)
Future development of intelligent where NTCIP will be deployed, the resulting di-
Transportation systems versity of application specific characteristics, such
as type and quantity of data to be transferred, the
After few decades’ efforts, ITS provides a lot of criticality of data transfer times, acceptable cost of
benefits to users, agencies, authorities, and the communications infrastructure, and the criticality
environment. There are still a lot of issues to be of data security and integrity issues.
solved and a lot of improvements to be made. Interoperability and interchangeability are the
Along with the development of new advanced key goals of the NTCIP open-standards effort.
technologies, ITS will provide us a better trans- The term interoperability refers to the ability for
portation environment that is also sustainable in multiple devices, often of different types, to work
the long run. together as a single system for certain common
purposes. For example, using the same commu-
nications channel to interconnect a management
nTcip system with traffic signal controllers, dynamic
message signs, video surveillance controls, and
NTCIP is different from the past practice of trans- other devices. The terms interchangeability gener-
portation management protocols in that it is not a ally refers to the ability to use multiple brands of
single communications protocol designed for one a device on the same communications channel, as
purpose. Rather, it is a whole family of protocols well as the ability to swap them out. For example,
covering from simple Point-to-Point protocols to the ability to put any brand of NTCIP-compatible
complicate objects oriented techniques. This is due traffic signal controller in the same system at the
to several reasons: the diversity of the applications same time.
41
communications standards the status of signals operated from another

for c2F and c2c center;
• A transit system reporting schedule adher-
NTCIP provides two different types of ITS com- ence exceptions, to a transit customer infor-
munications standards. The first type is between mation system or to a traveler information
an ITS management center and multiple control system, while also asking a traffic signal
or monitoring devices managed by that center. management system to instruct its signals
Below are some examples of this type of com- to a behind-schedule transit vehicle;
munications: • An emergency management system report-
ing an incident to a highway management
• A traffic signal management system com- system, to a traffic signal management sys-
municating with signal controllers at tem, and to a traveler information system;
intersections; and
• A traffic management system controlling • A highway management system inform-
dynamic message signs, advisory radio ing an emergency management system of
transmitters, and environmental sensors on a warning message in response to its notifi-
roadways; cation of an incident.
• A transit management system communi-
cating with passenger information signs on Such communication is called center-to-center,
transit vehicles and at transit stations and although two or more of the systems may be lo-
stops; cated within the same location or building. C2C
• A highway management system communi- involves peer-to-peer communications between
cating with detectors and ramp meters. any number of systems in an interconnected net-
work which is similar to the Internet where any
This type of communication is called center- center can request information from, or provide
to-field, since most applications of this type information to, any other centers. It is also pos-
involve a computer system at a management sible to use such protocols for communication
center communicating with various devices at to and between field devices, as well as between
the roadside or on vehicles. Protocols intended computer systems.
for these applications are often used in the en- Although both C2F and C2C communica-
vironment where a central management station tions can involve an operator making requests
routinely polls each field device, as in the most or issuing instructions, one of the features of the
common case of multiple field devices sharing a NTCIP protocols is their support for continuous
communications channel. and automatic functionality using pre-defined
The second type of communication involves data transmissions.
messages sent between two or more ITS manage-
ment centers. Below are some examples of this 3.2.1.1 Center-to-Field Protocols
type of communications:
NTCIP provides three application level protocols
• Two or more traffic signal management for C2F communications: the Internet’s Simple
centers exchanging information to achieve Network Management Protocol (SNMP) (IEEE
coordinated operation of traffic signals Std 828-1998, 1998; Feit et al., 1995; Stallings
managed by the different centers and to et al., 1996), the Simple Transportation Manage-
enable personnel at one center to monitor ment Protocol (STMP) and the Simple Fixed
42
Message Protocol (SFMP), which is currently closed circuit camera controllers. NTCIP
under development. is currently developing SFMP to meet this
need.
• Simple Network Management Protocol:
SNMP provides a simple, but bandwidth- These three protocols use the same get/set mes-
inefficient protocol for C2F applications, saging paradigm as that used in SNMP. Although
based on the Internet protocol of the same with the same base data elements, they differ in
name (SNMP). It is suitable only for net- the level of complexity to implement and the types
works with high bandwidth, or low volumes of services. Table 2 summarizes the comparisons
of messages. SNMP has been designed by among SNMP, STMP, and SFMP, while Figure
the Internet community to run over UDP/ 14 demonstrates the major advantages of these
IP, but it can be forced to run over TCP/IP protocols.
or T2/NULL. STMP is the most bandwidth efficient option
• Simple Transportation Management currently available and includes full support of
Protocol: STMP was developed specifi- SNMP for infrequent messaging demands. It
cally for use in the transportation industry. includes SNMP as a subset, so that any manage-
It is an extension of SNMP that allows C2F ment system that implements STMP can also
messages to be sent more efficiently using communicate with a device that supports SNMP. It
dynamic objects. Stacks based on this pro- also requires the use of SNMP to define dynamic
tocol are suitable for networks with low objects. Infrequent messages requiring additional
bandwidth and high volumes of messages, security can be sent using SNMP. STMP is the
including such traffic signal systems where most flexible and bandwidth efficient option. The
a central computer is directly connected to advantage of STMP is its support for dynamic
field devices, without the need to route the objects which are combined with a more efficient
information through some other devices encoding scheme, dramatically reduce the packet
such as an on-street master in a closed loop overhead relative to SNMP. Dynamic objects can
system. STMP has been designed to run also enable the user to define custom messages
over T2/Null since it supports low band- that are composed of any number of individual
width links, but could also be used over data elements. However, these data elements will
UDP/IP or TCP/IP if there is sufficient have to be defined in both the center and the field
bandwidth. devices in order to work properly.
• Simple Fixed Message Protocol: There The field devices that use any particular
is a need for having a bandwidth efficient subnetwork protocol can share the same com-
protocol for low-end field devices, like munications with other devices using the same
Table .2 SNMP, STMP and SFMP comparisons
SNMP STMP SFMP

Ease of implementation Easy Hard Currently Under
Development
Message set Support Limited to 13
Supports routing and dial-up Options Options
Can send any base data element? Yes Yes
Bandwidth efficiency-inverse of packet overhead Worst Best
43
Figure 14. C2F protocols

• Number of devices sharing the same line
or channel.
The first seven of these variables will determine

the total time needed to communicate with each
device. Although STMP is designed for use with
communications channels that use a slow transmis-
sion rate, it is not as bandwidth-efficient as most
proprietary protocols used in the past. With exist-
ing communications infrastructure, it may not be
possible to maintain the same polling period with
the same number of devices per channel. This is due
protocol. It does not matter whether such devices
to the fact that proprietary protocols are optimized
are from different manufacturers or are totally dif-
for each manufacturer’s equipment and consist
ferent devices, e.g., a light signal and a dynamic
of very few fixed short messages without any
message sign. Each device is assigned an address
flexibility in terms of changing these messages,
that is unique in that channel. The management
while standard protocols are flexibly designed
system can communicate with any of the devices
to accommodate all needs and a wide variety of
at any time by sending a message addressed to
information and messages in a multi-manufacturer
that device. However, when using Point to Multi-
environment. However, careful design can usu-
Point Protocol (PMPP), the management system
ally find a reasonable compromise between the
can communicate with only one of the devices
principal variables. Higher available bandwidth,
on the channel at a time. As a function of SNMP
bit rate, yields fewer compromises or required
and STMP, devices can only send a message to
trade-offs. If new communications infrastructure
the management system when requested by the
can be provided, it will allow for additional chan-
management system. The NTCIP protocols enable
nels and/or higher transmission rates.
broadcast messages for all devices, and no devices
can reply to a broadcast message.
Center-to-Center Protocols
When planning a C2F communications net-
work using NTCIP that involves continuous
NTCIP originally provided two application level
polling of field devices, e.g., a traffic signal
protocols for C2C communications, DATEX
system or transit fleet system, it is important to
(Data Exchange) and CORBA (Common Object
consider the relationship between the following
Requesting Broker Architecture), which were
key variables:
found necessary to meet the variety of require-
ments for inter-system data exchanges. Recently,
• Transmission schemes.
there has been increased interest in using XML
• Time between devices or between polling
and Web Service technologies for C2C links due
cycles.
to its simplicity and the wide accessibility of tools
• Transmission rate.
to provide these services. However, it is important
• Frequency of communication.
to determine where to deploy these protocols in
• Transmission delay.
the network, with some centers acting as bridges,
• Response delay in the field device.
or translators, among different protocols.
• Frequency of each type of message.
DATEX was designed to provide simple and
• Length of message to be sent.
cost-effective solutions for the following needs:
44
• Systems requiring real: time, fast data be informational or can constitute a command to
transfer; take some actions. Consider a message sent from
• Systems with limited communications one traffic signal system to another and contain-
bandwidth but high data transfer load; ing a signal timing pattern number. In DATEX,
• Systems with infrequent event driven ex- depending on the message type, it could represent
changes over dial-up links; and a command to implement that timing pattern at a
• Non-object oriented systems. particular traffic signal or group of signals, or it
could represent a status report indicating that this
DATEX provides a general-purpose C2C data timing pattern was just implemented at a particular
exchange protocol stack. It uses pre-defined mes- traffic signal or group of signals.
sages transmitted by the base Internet protocols The user can also establish standing subscrip-
(TCP/IP and UDP/IP) in a peer-to-peer network. tions for data. In DATEX, these subscriptions can
The base standard at the application level is an specify that data be sent one-time-only, periodi-
ISO standard. cally, or repeatedly on occurrence of some event
On the other hand, CORBA provides several as defined in the subscription. Each subscription
features to support networks connecting object message has a corresponding publication message.
oriented systems, and assuming sufficient pro- Unless the subscription is a one-time request, the
cessing power and communications bandwidth data will continue to be automatically published
are provided. Object oriented software can take repeatedly until the subscription is cancelled,
full advantage of CORBA and implement it eas- or until a predefined end date specified in the
ily; this is much more difficult to achieve with subscription.
traditional procedural software. A system can use CORBA to automatically
CORBA is a general purpose C2C communica- discover data availability and shared control op-
tions protocol based on the computing industry tions available from other systems. These other
standard of the same name. For object-oriented systems use the CORBA framework to publish
systems, it enables a higher degree of integration their capabilities and services offered, accept
and some services not provided by DATEX, but it registration requests from authorized clients, and
may not be suitable for near real-time applications then deliver those capabilities and services to
and loosely coupled systems. those clients on demand. For example, a CORBA
The wide availability of XML tools and large traffic management system that owns a CCTV
market have generated the market interest in XML. can offer to provide: (1) the images acquired as
It is especially suited for systems requiring lim- (a) snapshot, or (b) streaming video, and/or (2)
ited, simple data exchanges over communications allow remote control movement of that CCTV.
links with sufficient bandwidth and processors The system owning the CCTV is the server and
with sufficient processing time available. How- the system asking for the images, and/or control
ever, there are no current transportation industry of the CCTV is the client. This example also
standards for the use of XML. The NTCIP effort serves to illustrate a typical use of a subscription
continues to monitor the maturity of XML in an such as “send me a new snapshot image from
effort to determine its suitability for future use in CCTV every minute” stated in the proper terms
the transportation industry. for that CORBA system- assuming the requester
C2C networks allow each system to request is authorized that service, the expected result is
available information from any other systems. fairly obvious.
Each system can be configured to either accept C2C communications require a peer-to-peer
or reject certain requests. The data sent can either network connection such as a local area network
45
or a wide area network between the involved deployment should consider the impacts that this
centers. Local area networks typically use agency- may have on the long-term maintainability of a
owned twisted pair cable or fiber optic cable. system. The best solution is still likely to deploy
Wide area networks typically use commercial one of the recognized standards, but the agency
telecommunications links such as frame-relay, should realize that a future project would likely
fractional T1 leased lines, packet radio, leased be required to upgrade the software to address any
“virtual private networks”, ISDN, or similar mo- included features affected by revisions in order
dems over “plain-old telephone” lines. Any type to achieve the final mature standard.
of communication link can be used, as long as it
enables use of the Internet transport and routing The role of nTcip in the
protocols (TCP/IP and UDP/IP) and has sufficient iTs architecture
bandwidth for the planned communications load
to achieve the desired operational performance NTCIP defines a family of general-purpose com-
(this is based upon frequency, size of messages munications protocols and transportation specific
to be exchanged, and latency issues encountered data dictionaries/message sets (IEEE Std 1488-
when using C2C systems). 1999, 2000; IEEE Std 1489- 1999, 1999; IEEE
For DATEX and CORBA, the base protocols Std 1512-2000, 2000) that support most types of
have been defined, that is, how to exchange computer systems and field devices used in trans-
data, but the standards defining the data to be portation management. Applications for NTCIP
exchanged have not reached a state of maturity. are generally divided into two categories: C2F and
The XML approach is even less mature in that C2C. The former, normally involves devices at
the industry has not agreed on the exact rules on the roadside, communicating with management
how to exchange the XML documents. Any recent software on a central computer.
Figure 15. NTCIP and the ITS architecture
46
C2C applications usually involve computer-to- ◦ Vehicle detectors

computer communications where the computers ◦ Highway lighting control
can be in the same room, in management centers • Center-to-Center
operated by adjacent agencies, or across the ◦ Traffic management
country. The role of NTCIP in the National ITS ◦ Transit management
Architecture is illustrated in Figure 15. ◦ Incident management
For both C2F and C2C applications, NTCIP ◦ Emergency management
supports systems and devices used in traffic, tran- ◦ Parking management
sit, emergency management, traveler information ◦ Traveler information
and planning/data archiving systems. Figure 16 ◦ Commercial vehicle operations
illustrates how various transportation manage- regulation
ment systems and devices can be integrated using
NTCIP. Many applications of NTCIP are related to
The following are examples of systems and real-time communications and involve continu-
devices that can take advantage of NTCIP: ous transmissions of data or commands while
historical data can be sent using NTCIP, other
• Center-to-Field communication standards, especially e-mail and
◦ Traffic signals file transfer protocols developed for the Internet,
◦ Dynamic message signs may also be used. Human-to-human communica-
◦ On-board sensors and controllers tions are generally better served by fax/telephone
◦ Environmental sensors and Internet protocols, but basic support is also
◦ Ramp meters provided in the NTCIP C2C protocols.
Figure 16. Example of ITS integration using NTCIP
47
The nTcip Framework as part of a protocol stack. There is also another

way to classify the NTCIP standards: primary,
When options are available with layered protocols, supporting, and base standards protocols.
the options can be diagramed in a “framework.”
Figure 17 illustrates the framework for NTCIP. The • A primary standard applies directly and
diagram shows the different protocols that can be specifically to the device or component
chosen at each level and which ones are compat- subsystem being implemented. For exam-
ible. However, not all compatible configurations ple, standards 1203 and 1204 are primary
make sense, and there are mutually exclusive NTCIP standards and the Standard 1203
choices. A particular message transmission can applies specifically to DMS, and 1204 to
use at least one protocol from each level of the ESS.
NTCIP framework. The series of protocols used • A supporting standard applies in general
in the message transmission is called a “protocol to more than one device or component
stack.” It is possible for a pair of electronic devices subsystem implementation. For example,
to exchange some messages using one stack and the NTCIP Standard 1201 Global Objects
other messages using a different stack, though standard applies to all devices and compo-
usually, such stacks will differ only at one or nent subsystem implementations that use
two levels or sublevels. As shown in Figure 17, or require features such as identification
the lines connecting standards at different levels and location of equipment, global time,
show optional standards at each level. If there is and event detection or scheduling. Thus,
a continuous line from one standard to another, standard like Standard 1201 is a support-
then they are compatible and can be used together ing standard. Both primary and supporting
Figure 17. NTCIP standards frameworks
48
standards typically apply at the C2F or protocols generally offer further options within
C2C information layer (Figure 17). most of the standards. Examples of sub-options
• A base standard and protocol applies to the within a standard are: which subset of messages
Application, Transport and Sub-network are supported, or which bit rate is used at the
levels. These standards define NTCIP physical interface.
unique capabilities for protocol and data
transport choices to complete the design nTcip communication levels
of an operational deployment. These stan-
dards is different from both primary and NTCIP uses a layered approach to communications
supporting standards, since the data being standards, similar to the layering approach adopted
exchanged is irrelevant. These standards by the Internet and the International Organization
are unaware and largely unaffected by of Standards (ISO). In general, data communica-
their use in a signal control, DMS, and ESS tions between two computer systems or other
applications. electronic devices can be considered to involve
the following primary layers, called “levels” in
When deploying an NTCIP-based system, NTCIP, to distinguish them from those defined
protocols have to be chosen. Figure 18 illustrates by ISO and the Internet. The NTCIP standards
an example if C2F protocols stack choice that publication numbers are grouped in number ranges
can be defined using NTCIP standards. A stack to indicate the standard type and the level where
is a subset of the overall NTCIP framework-a se- the standard goes.
lected route through the levels, given the choices
available. Some stacks include two standards at • Information level: This level contains
some levels, which usually mean the protocol standards for the data elements, objects
can use either of the optional standards. NTCIP and messages to be transmitted, e.g., TCIP,
Figure 18. Example center-to-field stack
49
NTCIP 1200 series Standards Publications, these areas. The two major exceptions are the
MS/ETMCC. protocols that support:
• Application level: This level contains
standards for the data packet structure and • Slow speed, high frequency communica-
session management, e.g., SNMP, STMP, tions links as found in 1200 bps, once-per-
DATEX-ASN, CORBA, FTP. second traffic signal systems.
• Transport level: This level contains stan- • A simplified Publish-Subscribe C2C
dards for data flow control, packet reas- protocol.
sembly and routing when needed, e.g., • NTCIP has extended existing standards or
TCP, UDP, IP. developed entirely new protocols as need-
• Subnetwork level: This level contains ed in cases where ITS has special protocol
standards for the physical interface, e.g., requirements. The two areas include:
modem, CSU/DSU, and the data frame • Continuous, automated, real-time exchange
encapsulation method, e.g., HDLC, PPP, of large volumes of small data packets in a
Ethernet, ATM. many-to-many multi-agency network.
• Plant level: This level consists of the • Continuous high volumes of real: time data
physical transmission media used for com- sent to and from embedded processors in
munications, e.g., twisted pair copper wire, roadside or on-vehicle equipment sharing
coaxial cable, fiber optic cable, wireless. It the same, low-speed, data channel and re-
should be noted that the plant level is an quiring low latency.
infrastructure choice and not a standards
selection choice. However, the plant level Through a layered combination of existing
selection will have an impact on the sub- communications standards and a few new stan-
network level selection to which it must dards developed specifically for ITS, NTCIP
interface. provides a family of communications protocols
that serve many of the common needs in ITS
Information level standards used in ITS are transportation management.
unique to the transportation industry. The National The levels shown in the framework are some-
ITS Architecture and much of the standards de- what different from communication stack layers
velopment effort for ITS involve identification of defined by the ISO’s Open Systems Interconnect
required data elements and the definition of their seven-layer reference model and other standards
use for all the different domains and functions developing organizations. The NTCIP stack
within ITS, e.g., traffic, transit, traveler informa- extends beyond the communications stack to
tion, emergency management. At the Application, include informational data and interfaces to the
Transport and Subnetwork levels, ITS can fre- actual communications infrastructure. The levels
quently use existing standards used by the broader and terminology used in NTCIP were chosen for
computer and telecommunications industries. simplicity and ease of understanding by readers,
Below the Information level, the NTCIP standards and related to typical applications in the trans-
deal with choosing which existing standards are to portation industry.
be used in ITS. The Internet standards have been With the many diverse requirements of NTCIP,
adopted where possible. The NTCIP standards it is not surprising that we looked at the ISO OSI
specify which options to use where alternatives Reference model to help us define the framework
are available in some standards. NTCIP has not for the new family of standards. Although OSI
had to develop significantly new standards in communications protocols are not widely used, the
50
Figure 19. Level mapping of OSI layer to NTCIP level mapping
layered model remains. The OSI model breaks the defining the rules for greeting each other
communications process into seven well-defined and exchanging information. These stan-
layers. Each layer has a defined purpose, generally dards are equal to the Session, Presentation
independent of adjacent layers. Figure 19 shows and Application Layers of the OSI model.
how the NTCIP Information, Application, Trans- • NTCIP transport level: This level defines
port, Subnetwork and Plant Levels loosely relate the rules and procedures for exchanging
to the OSI model and are described as follows: the Application data between source and
destination on a network, including any
• NTCIP information level: This level de- necessary routing, message disassembly/
fines the meaning of data and messages re-assembly and network management
and generally deal with ITS information. functions (Buede et al., 2000). This is
This is similar to defining a dictionary and similar to the rules and procedures used by
phrase list within a language. These stan- the telephone company to connect two re-
dards are above the traditional OSI seven- motely located telephones. Transportation
layer model. Information level statndards level standards are roughly equivalent to
represent the functionality of the system to the Transport and Network Layers of the
be implemented. OSI model.
• NTCIP application level: This level • NTCIP subnetwork level: This level de-
defines the rules and procedures for ex- fines the rules and procedures for exchang-
changing information data. The rules may ing data between two devices over some
include definitions of proper grammar and communications media. This is equivalent
syntax of a single statement, as well as the to the rules used by the telephone company
sequence of allowed statements. This is to exchange data over a cellular link ver-
similar to combining words and phrases to sus the rules used to exchange data over a
form a sentence, or a complete thought, and twisted pair copper wire. These standards
51
are roughly equivalent to the Data Link • Point-to-Point Protocol (PPP)

and Physical Layers of the OSI model. • A customization of the High-level Data
• NTCIP plant level: This level is used to Link Control (HDLC) standard at the
provide a point of reference to those learn- subnetwork level, known as the Point-to-
ing about NTCIP. The Plant level includes MultiPoint Protocol and
the communications infrastructure over • SNMP standard at the application level.
which NTCIP communications standards
are to be used and will have a direct im- Each standard specifies one or more protocols
pact on the selection of an appropriate to be used at a given level, and the suboptions al-
Subnetwork Level for use over the se- lowed or required within each of those standards.
lected communications infrastructure. The A standard publication will typically reference one
NTCIP standards do not prescribe any one or more “normative standard” publications-other
media type over another. To ensure a work- publications that contain additional relevant speci-
ing system, deployers must specify and/or fications for the standard(s). A referred normative
select an NTCIP protocol or profile at each standard may be another NTCIP publication, if the
level. standard was developed by NTCIP, or a publica-
tion developed by any other standards develop-
Most of the standards in the lower levels are ment organization. For a particular application of
existing commercially available standards used NTCIP, the user must select, in the procurement
in the telecommunications industry and were not specifications, which elements are desired at each
developed uniquely by NTCIP, although NTCIP level. e.g., select from the options called out in one
often specifies which sub-options within those or more profile publications for each level. The
standards are to be used. The majority of standards set of selections and options for base standards
unique to Intelligent Transportation Systems are and protocols for all levels is referred to here as
found in the Information Level and Application a “protocol stack.” Each NTCIP protocol stack
Level shown at the top of Figure 17 and Figure 18. will have different characteristics, and a stack that
Each NTCIP protocol stack involves a mixture of works well for one application or communications
standards, with at least one from each level. environment may not suit another.
Application level standards that NTCIP has
major protocols in nTcip defined are briefly described below. As pointed
protocol stacks out in Fig. 3.17 and 3.18, these application profiles
can be combined with certain transport profiles.
The first NTCIP standards developed were those Two electronic devices will be better able to
intended for C2F applications. This involved communicate interchangeably with each other, if
a new application level standard called Simple they use the same communications protocol stack,
Transportation Management Protocol, a new the same information level data dictionaries and
transport level standard called the Transportation message sets, and implement the same desired
Transport Profile (T2 or T2/NULL), and several options defined in each of these selected primary,
sets of new standard data elements (Michael et al., supporting, and base standards and protocols.
2004; IEEE Std 1489-1999, 1999) called “object In addition to specifying a protocol stack,
definitions” at the information level. The initial the system designer must also choose between
NTCIP C2F protocol development also involved various options and alternatives available in the
references to three existing standards: selected stack. These options exist in both C2C
and C2F protocol stacks. Major options, such as
52
which protocols to support at each level in the plish these migration steps. Pursuit of a migration
communications stack, are sometimes grouped strategy towards the use of open standards starts to
according to conformance levels, while others minimize the use of proprietary communications
are individually selectable. Most manufacturers and begins to maximize the use of NTCIP. Finally,
and system suppliers typically offer features that at some future point, the migration is completed
go beyond the standard. To make use of such and NTCIP is fully deployed, having replaced all
features, it is necessary to specify the inclusion of now retired legacy systems.
manufacturer-specific data elements or messages NTCIP and non-NTCIP devices may be mixed
as extensions of the standards when procuring a on the same channel. Thus, all devices sharing
management system. a channel must be upgraded simultaneously. A
The decision by an agency to use features center that communicates with both NTCIP and
above and beyond the standard should be taken non- NTCIP devices will need to use a different
only with the understanding of the potential im- communications port for NTCIP devices and for
pacts. These impacts could be considerable in the non-NTCIP devices, and will need to support both
long term. These options may, in effect, result in protocols. Therefore, the mixed devices listening
the purchase of proprietary systems. Part of the on the shared communications channels must
decision must include how many of these features recognize and react only to those data elements
that will be allowed. and commands intended for them individually, and
must also not produce unpredictable results in re-
migration from legacy sponse to any other data traffic on the channel.
systems to nTcip For example, the most likely and simplest so-
lution in the traditional closed-loop traffic signal
Since interoperability and interchangeability are systems is to limit each field master to one pro-
two key goads of NTCIP, the inability to update tocol. Only field masters with NTCIP-compatible
older equipment should never stop an agency controllers would be upgraded to support NTCIP.
from replacement or migration strategies to make This avoids the need for field masters to simul-
full use the benefits of NTCIP conformant imple- taneously support two protocols on two separate
mentations. For example, a central system whose ports. The center could communicate with field
current field devices cannot be updated could be masters using a different protocol than that used by
expanded to run NTCIP protocols on some com- the field master to communicate with controllers.
munications channels while the older equipment As with the controllers and the field master, the
is maintained on others. center’s software will need to be modified to add
There is a model for a three-step migration support for an NTCIP protocol, if NTCIP is to be
from legacy systems to NTCIP. Initially, the used for communications with field masters.
proprietary interface details may or may not be Any upgrade of an existing system to add sup-
known. Then, there is some intermediate state port for NTCIP should be designed in consultation
and some period of time where the operational with the system provider. Each provider should
system consists of a mixture of the legacy systems adopt an upgrade or migration strategy that is
and the newer NTCIP hardware. There may be most efficient for the majority of its custom-
a common communications channel for legacy ers. If a customer wants a unique arrangement,
and NTCIP devices. The central control system that customer may have to pay the full cost of
may be separate or combined; it may run on the the software modifications, whereas the cost of
same computer or on separate computers: this is the general solution can be spread among many
determined by the scope of the project to accom- customers. One approach to the introduction of
53
NTCIP in a C2F system is to operate NTCIP and Specifically, this amendment will address Agency
non-NTCIP systems during a transition period. issues concerning current implementation efforts
Field devices can gradually be switched over and their needs based upon experience gained
from one to the other as they are replaced or their through NTCIP deployment experience.
software is upgraded. If the current system is quite The Washington State Department of Trans-
old and upgrading it for NTCIP is not practical, portation (WSDOT), USA implemented a Vari-
this transition should be done as part of a general able Message Sign (VMS) system in 1999. The
system upgrade. formal name of this project is the “NTCIP VMS
Software Upgrade.” The purpose of the project
nTcip implementation examples was to modify the existing traffic management
system to support selected protocols from the
As shown in Figure 20, NTCIP have been de- NTCIP protocols and to purchase two new NTCIP
ployed across U.S.A in several states. The initial compliant variable message signs. The vendor
deployment of NTCIP-conformant equipments was American Electronic Sign Co (AES). Ad-
was conducted by the Virginia Department of ditionally, the AGENCY hired a programming
Transportation (VDOT), USA in a case study of contractor (PROGRAMMER) to enhance the
their Variable Message Sign implementation in central system software.
NTCIP-9002 Version 01.04, in September, 1999. The City of Phoenix, Arizona, USA initiated
AASHTO, FHWA, ITE and NEMA are currently a project to enhance their traffic signal system as
sponsoring this case study update. This effort, shown in Figure 20. This project included two
presented as a case study amendment, focuses distinct parts, the replacement of the central traffic
on insights gained over the three years of deploy- control system, and the upgrade and purchase of
ment since the initial case study was performed. additional traffic signal controllers. The NTCIP
Figure 20. NTCIP projects in U.S.A.
54
was specified as the communications protocol of using SNMP, the STMP communication is sug-
choice for both. The central system, called Phoe- gested to be used in England.
nix ATMS was also required to control additional
field devices, however, without the requirement to
utilize the NTCIP communications protocol. reFerences
The City of Lakewood, Colorado, USA initi-
ated a project to enhance its traffic signal system. Buede, D. M. (2000). The engineering and design
This project included two distinct parts, the re- of systems: Models and methods. New York:
placement of the central traffic control system, Wiley, Inc.
and the upgrade and purchase of additional traffic Feit, S. (1993). TCP/IP: Architecture, protocols
signal controllers. The NTCIP was specified as the and implementation. New York: McGraw Hill,
communications protocol of choice for both. The Inc.
central system, called Lakewood ATMS was also
required to be extensible in order to add future Feit, S. (1995). SNMP: A guide to network man-
capabilities for controlling field devices such as agement. New York: McGraw Hill, Inc.
NTCIP Dynamic Message Signs and non-NTCIP
Institute of Transportation Engineers Manage-
closed circuit television cameras.
ment and Operations of Intelligent Transportation
The City of Mesa, AZ initiated a Request for
Systems. (2003). ITS standards overview.
Proposal to upgrade their existing SONEX sys-
tem, which included a requirement for NTCIP, Internet Librarian, D. I. S. A. (1997). US-DOD
i.e., “Support NTCIP for communication to TS-2 Internet related standardized profiles. Retrieved
controllers”. However, the proposed cost was from http://www-library .itsi.disa.mil/
so much higher than estimated, that the Agency
Michael, A. (2004). Guide to the IEEE 1512
withdrew the RFP. In 1997, the Agency again
family of standards. Washington, DC: Institute
issued a Request for Proposal, this time request-
of Electrical & Electronics Engineer.
ing to replace the entire signal system including
the central system and the signal controllers. A Research and Innovative Technology Administra-
phased approach was to be used running temporary tion (RITA) & U.S. Department of Transportation.
parallel central systems. The NTCIP was speci- (US DOT). (2002). National ITS architecture.
fied as the communications protocol of choice Retrieved from http://itsarch.iteris.com/itsarch/
for the new components. The central system,
Stallings, W. (1996). SNMP, SNMPv2 and RMON.
called Mesa ATMS was also required to control
Reading, MA: Addison-Wesley Publishing Com-
additional field devices and provide an interface
pany, Inc.f
to the SYNCHRO software.
In England, a project, called UTMC (Urban NTCIP Standard 9001. (2002). The NTCIP
Traffic Management and Control) had been guide.
implemented and evaluated. However, the NTCIP
ITS Standards Outreach, Education and Training
definition cannot support the England metropolis
Program, Institute of Transportation Engineers.
area at present the material transmission demand.
(2006). Center to center communications.
If NTCIP is applied to the European area, the MIB
file must be revised or established to conform IEEE Std 828-1998. (1998). IEEE standard for
to the application demands of the England area. software configuration management plans.
UTMC indentified that SNMP communications
produce serious overhead problem. Instead of
55
IEEE Std 1489-1999. (1999). IEEE standard for IEEE Std 1512-2000. (2000). IEEE common
data dictionaries for intelligent transportation incident management message sets for use by
systems. emergency management centers.
IEEE Std 1488-1999. (2000). IEEE standard for
message set template for intelligent transporta-
tion systems.
56
Section 2
Embedded System Architecture
and Communication Protocols
58
Chapter 4
Vehicular Embedded
System Architecture
Chung-Ping Young
National Cheng Kung University, Taiwan, R.O.C.
absTracT
The dramatic advancement of IC technologies makes electronic devices be smaller and run faster, so
they are able to implement more functions in a limited space. The car electronics play an increasingly
important role in automobile industry, and the embedded system has already been extensively employed
for improving the operation and performance of vehicles, such as safety, comfort, convenience, and
energy consumption. In terms of electronic system, an automobile is a distributed embedded system,
and the control messages to each electronic control unit (ECU), go through in-vehicle networks. An
ECU is a computing system, integrated with a data acquisition module or an electromechanical driver.
A variety of ECUs implement versatile functions, such as powertrain, antilock braking system (ABS),
traction control system (TCS), adaptive cruise control (ACC), and electronic stability program (ESP),
etc. Sensors provide measurements of specific vehicle parameters in a format suitable for the digital
microcontroller, while actuators are electrically operated devices that drive electromechanical compo-
nents. Human machine interface is the input and output of vehicle operations to users.
inTroducTion once required and in addition largely extended

the living environment that people can reach. The
Transportation of humans and objects have been more time and labor for transportation is saved, the
playing an important role in our daily lives since more leisure time people will have. Animal-power
civilization first formed and needed new means of or natural resources have been the driving force
reaching destinations. The invention of efficient of transportation for a long time. After the steam
transportation greatly reduced the time and labor engine was invented, the automobile started a new
era. The mass production of the Ford model T cre-
DOI: 10.4018/978-1-60566-840-6.ch004
Vehicular Embedded System Architecture
ated the modern automobile industry and made systems are employed for data processing and
the automobile more affordable. control in an automobile. The dramatic advance-
The basic structure of the automobile has not ment of IC technology, which is described by
changed much, but evolving technologies has kept Moore’s law, makes chips smaller, faster, and
improving its functions and performance. The con- able to implement more functions. In terms of
struction of traffic networks and mass production electronic systems, an automobile is a distributed
of automobiles have made the automobile the most embedded system, and the control messages to
important land based transportation carrier. The each distributed device go through the in-vehicle
usage of automobiles is usually associated with network. X-by-wire is becoming a new technical
the growth of economy and industry of a nation, trend.
so the population ratio that owns automobiles in a From a top-down viewpoint, the ultimate
developed country is larger than that in a develop- goal of transportation is to develop an intelligent
ing country. When the economy grows, vehicle transportation system. The basic mobile unit is a
as a transportation tool becomes more affordable vehicle, which is interconnected to other vehicles
and popular, for instance China or India. When or backend service providers through vehicle-to-
people use automobiles in their daily lives, they vehicle or vehicle-to-infrastructure communica-
demand not only mobility, but also safety, comfort tions. The scope of this chapter is limited to the
and convenience. These are some design factors distributed embedded systems in a vehicle.
that manufacturers have to put into aspect when A vehicle consists of a variety of electronic
enhancing functions by introducing and develop- control units interconnected through an in-vehicle
ing new technologies. network, while each unit is an embedded system
For a government to provide a modern trans- involving processor and memory along with other
portation system, it has to build not only a traffic optional sensors, actuators, storage devices or
network, but also an infrastructure to access more human machine interface.
information to allow drivers and passengers to
drive safer, more comfortably and with better
convenience on the road. This is the vision of an saFeTy, comForT and
intelligent transportation system. To achieve this conVenience
goal, both the infrastructure and vehicles have
to be facilitated with a modern electronic and Vehicles were developed for transportation.
information system. When vehicles become mandatory transporta-
The evolution of an automobile shows more tion tools in daily life, safety is the first issue.
signs of adopting electronic devices. To enhance Road safety is related to the loss of human life
the features or performance, some mechanical and property and can be categorized into three
components are replaced by wires and electronic areas: human, environment, and vehicle. Human
devices, or simple electronic devices are enhanced and environment factors are out of the scope, and
by complex electronic control systems. An auto- we will focus only on the vehicle. However, the
mobile consists of several control systems: power enhancement for vehicle safety can sometimes
train, chassis, safety, body and information. Each compensate the inappropriate operation caused by
system may have several subsystems distributed in human or environmental factors. Vehicle safety
different location of a car and are linked through can be further separated into active safety and
in-vehicle networks. passive safety (Robert Bosch, 2006). The active
Since most modern electronic devices digitize safety mechanism is to prevent the happening
signals and process them by software, embedded of potential accidents, while passive safety is to
59
help drivers and passengers lower injury or death crash stages and the active/passive safety related
rates by accidents. Figure 1 shows the different safety mechanism.
categories of road safety, and goals for active The lifetime of market penetration of one auto-
safety and passive safety. motive system usually begins with an innovative
Active safety systems, including antilock brak- technology applied to a high-end model vehicle.
ing system (ABS), traction control system, and Later on this technology is recognized as a major
electronic stability program, enhance the stabil- improvement to environment or safety and re-
ity and steerability of driving so that corrections quired by law, or recognized as a mandatory feature
or reductions toward inappropriate operations by the customer. This product then becomes the
caused by humans or the environment can be standard equipment in all car classes. Seat belts,
made, therefore improving road safety. Adap- airbags, and tire pressure detection are regulated
tive cruise control not only provides the safety by legislation to be equipped on a vehicle, while
function of maintaining safe headway with the blind spot detection and adaptive front lighting
preceding vehicle, but also relieves the drivers are trying to be regulated in Europe.
need to frequently check the speed. The comfort and convenience are not as criti-
A car accident happens through three phases: cal as safety for evaluating a vehicle. To provide
pre-crash, in-crash and post-crash stages. To a joyful and easy driving environment, there are
prevent the happening of a tragedy, the safety important design factors to be considered when
mechanism must successfully work before the increasing the quality of driving and riding. Some-
pre-crash stage. The active safety provides safe times, some features of comfort and convenience
and convenient drivability to reduce accidents, are also correlated with safety improvements. For
and the passive safety provides pre-triggering example, blind-spot warning produces a visual/
to the protection device for the impact and post- audible signal to notify the driver of potential
crash communications. Figure 2 presents three dangers: a car approaching in the blind-spot area.
Figure 1. Road safety categories
60
Figure 2. Active and passive safety in different crash stages
The driver can concentrate on the front view of central console sends the command to ECU and the
the forward direction, but not be distracted by the ECU sends the status back to central console.
vehicle on either side. A vehicle can be categorized by the following
systems on vehicular network: powertrain, chassis,
safety, body, central console, and infotainment.
neTworked embedded sysTems Figure 3 shows that all vehicle electronic con-
trol modules are networked through in-vehicle
overview network.
To fulfill a variety of functions in a vehicle, each powertrain

electronic control unit (ECU) independently
implements its designated function. The auto- The powertrain system consists of engine control,
motive system is constructed as a server-client automatic transmissions, traction control system
architecture. ECUs can be treated as a peripheral (TCS) and an adaptive cruise control (ACC). The
device of the central console, and the digital data engine control is to gain higher engine power at
are communicated via the in-vehicle network. The a lower system cost level, while also offering
improved fuel efficiency and reduced exhaust
Figure 3. Networked vehicle electronic control modules
61
emissions. Transmissions convert the torque as power windows, power sliding/rear doors, door
generated by the engine and the engine speed lock, wiper, ventilation, heater, air-conditioning,
corresponding to the tractive force requirement. room lamp, and adaptive front lighting.
The automatic transmissions automatically take
on control of starting, selecting the gear ratios and central console
switching gears. The TCS prevents the wheels
from spinning by reducing the drive torque at Central console is the user interface to the automo-
each driven wheel. The ACC system maintains the tive system. It involves dashboard projection of
constant speed set by the driver, and will reduce the vehicle status and button/knob to control the
the speed to follow a slower vehicle. It speeds up operation and setting. A concept central console
by electronically accelerating up to the setting via will have only a graphical LCD display with
the engine management system or decelerates by touch screen panel integrated. The communica-
electronically activating the brake system. tion between the central console and other ECUs
relies on the in-vehicle network, such as CAN/
chassis LIN, FlexRay, and MOST (Paret, 2007).
Chassis systems enhance vehicle stability and infotainment

steerability. Electronic stability program (ESP)
improves vehicle handling and braking response Infotainment system provides the contents of
through programmed intervention in the braking information and entertainment to occupants. En-
system and/or powertrain. It integrates both ABS, tertainment involves audio, video, radio, digital
preventing from wheel locking when brakes are TV, or gaming. The infotainment contents can
applied; and TCS, preventing from wheel spin- be retrieved from local mass storage devices or
ning when acceleration is applied. The suspension via radio, mobile communication, or wireless
adjustments compensate the rise and fall of wheels network to remote service/content providers. The
on cornering or road irregularities to optimize navigation system provides location awareness
vehicle stability. and route planning functions by utilizing multiple
xGPS systems.
safety
X-by-wire
When the automotive system senses an impact,
the passive safety systems activate. Front or side Since each ECU is located separately and con-
airbags are triggered and fully inflated in about 40 nected through in-vehicle networking, the con-
ms and vehicle occupant detection system drives trol of an electromechanical device is not by the
the seat belt pretensioner to function in restraining conventional mechanical parts, but by the digital
the occupants in their seats. The ABS prevents control message on the bus. X-by-wire implies
the wheels from locking up when the brakes are that the operation on a vehicle is implemented
applied by lowering the wheel brake pressures. through electrical wires, where x could be any
operation that you name. This reduces the cost and
body and comfort control space for mechanical parts, and the automotive
system appears as a mobile vehicular network.
Body and comfort control systems provide oc- Some examples of x-by-wire are steer-by-wire
cupants convenience and comfort features, such and brake-by-wire (Paret, 2007).
62
elecTronic conTrol uniT ECU is modeled as a closed-loop control system.

The ECU performs a looped control by sensing
overview the automotive system parameters, processing
the computation for determination of appropriate
During the technical evolution of automobile reaction, and then outputting the control signal to
development, the relatively straightforward adjust the operation of the automotive system. This
electromechanical operations in a vehicle cannot control loop starts again, when the automotive
fulfill the complex safety or comfort requirements system parameters change (Robert Bosch, 2006).
in a modern vehicle. The electronic devices have Figure 4 demonstrates the closed-loop control
become more popular in the 1960s, since the dawn of automobile steering under variable human,
of the semiconductor era. The computing systems vehicle, or environmental parameters.
brought signal processing to a brand new digital Some examples of ECU implementations are
world and then enhanced the system control. described in the next few subsections. Due to the
Because of the advancement of semiconductor strict operation environment in a vehicle, some
technology, the market pushed the automotive factors are carefully considered before develop-
systems to a more precise and complicated control ing an ECU.
in fuel injection, braking, and traction control.
Electronic Control Unit (ECU) is basically a • Size: Though the dimensions of a vehicle
processor or computing system, integrated with seem to be large, it is designed to have most
a data acquisition device or an electromechanical of the space reserved for driver and passen-
driver. Sensors provide measurements of specific gers. The physical size of an ECU has to
vehicle parameters in a format suitable for the be small and is integrated into the electro-
digital microcontroller. Similarly, actuators are mechanical module. The advancement of
electrically operated devices that regulate signals the semiconductor manufacturing process
to the device that directly control its output. The
Figure 4. Closed-loop control of automobile steering under variable human, vehicle and environment
parameters
63
makes this miniaturization requirement performance is required to be efficient to

feasible. ensure the appropriate control.
• Cost: Cost is always an important factor • Stability: The implementation of an ECU
in promoting a product on the market. The does not only require the performance but
electronic components are utilized more also the accuracy. This is to ensure the ap-
than ever in a modern vehicle where sev- propriate functioning of the system. The
eral ECUs can be utilized in an automotive conventional mechanical control can be
system, so the cost of the electronic devic- investigated by vision or instrumentation,
es is also a concern. All in all, processor but the operation of an electronic system
technology, IC technology, or component with hardware and software is invisible and
specifications determine the cost. has to be tested through a suite of carefully
• Function: Due to different requirements defined test cases. However, this still does
for individual subsystems in an automotive not guarantee the perfect functions without
system, ECUs are developed for various any errors. The fault tolerant or self-recov-
automotive functions. A dedicated com- ery mechanism must be considered to be
puting system embedded in an ECU can implemented in the system to enhance the
optimize its control performance. Though stability.
the system architecture is similar, the hard- • Robustness: Due to severe operation envi-
ware of sensors, actuator components, pro- ronments in a vehicle, if the hardware is not
cessor technology, and firmware of data robust enough it may cause the software to
acquisition, control and signal processing malfunction. The electronic components
are different. and the board design have to be developed
• Performance: Powertrain or brake con- on a stricter industrial standard for tolerat-
trols are time-critical when driving on the ing temperature, humidity, electrical noise,
road. A real-time operation, which guaran- or electromagnetic interference.
tees the designated tasks can be finished
before the critical deadline, is required. Figure 5 depicts the block diagram of an elec-
Vehicle speed is very high on highway and tronic control unit, while the blocks surrounding
human reaction time is short, so the system the computing platform are optional depending
on the function of the ECU.
Figure 5. A complete block diagram of electronic control unit
64
powertrain Figure 6 shows the simplified block diagram of

an antilock braking system.
Powertrain management controls the engine and
transmission. From the sensor inputs, the environ- Traction control
mental information, driving-conditions, vehicle
and user parameters are obtained for the powertain When the driver presses the accelerator, the
controller to determine the operation of the engine engine torque is converted to drive axle torque,
and transmission; then the powertrain controller and then is distributed to driven wheels via the
commands the engine, traction controller, and transversal differential. If the increased torque can
transmission to generate the appropriate torque, be transferred completely to the road surface, the
slip, or gear ratio (Robert Bosch, 2006). vehicle can stably accelerate. In some low friction
coefficient road surface situation, when the drive
antilock braking system torque can not be transferred completely, the wheel
will spin and the vehicle becomes unstable due
The ABS control system includes ABS ECU, to the loss of lateral stability.
wheel module, brake pedal module, and hydraulic The traction control system (TCS) ECU and
modulator (Robert Bosch, 2006). When force is transversal differential lock controller regulates the
applied on the brake pedal, the braking operation slip of the driven wheel by determining a reference
works through: brake booster, master cylinder, slip value, sensing the differential wheel speed,
hydraulic modulator, wheel-brake cylinder and calculating distribution ratio of the transversal
then brake. The ECU monitors the wheel speed; differential, and correcting the drive torque and
the master cylinder, pressure, calculates the ac- braking torque (Robert Bosch, 2006). The drive
tual and required slip and regulates the hydraulic torque can be adjusted by engine intervention,
modulator. The hydraulic modulator has solenoid while the braking torque regulated for each wheel
valves, which adjust pressure from the master is via the expanded ABS hydraulic system. Figure
cylinder to the wheel-brake cylinder. The ABS, 7 depicts the block diagram of a traction control
preventing the wheels from locking up, works system.
when the wheel brake pressures are lowered.
adaptive cruise control
The ACC controller includes cruise control, track-

Figure 6. Block diagram of antilock braking ing control, and acceleration control (Robert Bo-
system sch, 2006). The cruise control adjusts the vehicle
speed until it is equal to the set speed. The tracking
control selects one of the vehicles ahead as the
following target, and calculates the acceleration
based on the distance and relative speed. The
acceleration control sends the acceleration com-
mands to the powertrain or braking system. Figure
8 demonstrates the functional block diagram of
an adaptive cruise control system.
65
Figure 7. Block diagram of traction control system
Figure 8. Block diagram of adaptive cruise control system
electronic stability Figure 9 shows the simplified block diagram of

an ESP control loop. The sensors involve yaw-rate,
The electronic stability program (ESP) improves steering wheel-angle, pressure, and wheel-speed
vehicle handling and braking response through sensors, while the actuators include hydraulic
programmed intervention in the braking system modulators for braking and engine management
and/or powertrain to prevent the linear, lateral, or ECU for drive torque.
yaw velocity from exceeding control limits (Robert
Bosch, 2006). The ESP integrates the capabilities
of both ABS and TCS into a closed-loop control, processors
and provides each wheel highly precise perfor-
mance of the dynamic braking and longitudinal A processor is a semiconductor chip with digital
and lateral forces under various circumstances. circuits, which performs computational or logi-
cal tasks. It consists of three parts: the controller,
66
Figure 9. Block diagram of electronic stability

prototyping. The semicustom application-specific
program system
IC (ASIC), like gate array or standard-cell ASIC,
compromises these two IC technologies mentioned
above, so part of the arrays of gates or logic-level
cells have already been built and system developers
just need to connect these predefined components
to implement their design (Vahid et al., 2002).
According to the arrangement of the con-
troller, datapath, and memory in a processor;
processor technology can be categorized into the
following three types: general-purpose processor,
application-specific processor, and single-purpose
processors. The processor technology is relevant
to the hardware and software design of an embed-
ded system, so the application is critical and they
are described below.
general-purpose microcontroller
datapath, and memory. The controller decodes
General-purpose processor or microprocessors
the logic of software instructions or changes the
are designed for almost all of the common ap-
transfer among the hardware state machines. Data-
plications. The logic is implemented by software
path is the component executing computation or
programming, which has the maximum flexibility
logic functions for the specified data. Memory is
to alter its implementation, but has the lowest cost
the storage space for data and program. Different
for modification. A microcontroller is a micro-
processor technologies emphasize differently on
processor along with some peripheral modules,
the structure of these three processor components,
like memory and I/O, integrated in a single chip
so they are designed for different system archi-
without external circuits, so it provides extensive
tecture development.
functions and is widely used in many embedded
The processor is manufactured as an integrated
control systems. A more complicated system-on-
circuit (IC) component. It is implemented in dif-
chip (SoC) technology integrates more peripheral
ferent IC technologies, which determine the level
modules in one chip, almost a complete system
that the chip is customized. A full-custom/very-
even including analog circuits. So the SoC micro-
large-scale integration (VLSI) technology opti-
controller is able to implement more sophisticated
mizes performance and functions of a processor
functions to act as a full system.
from all aspects of digital circuit design. The
Among the processor technologies, this is the
manufacturing of a full-custom process has high
most common approach. The system implemented
nonrecurring engineering (NRE) costs and long
on a general-purpose microprocessor is referred
turnaround times. On the other hand, program-
to as a software approach. According to differ-
mable logic device (PLD) is off-the-shelf and is
ent applications, many manufacturers or product
ready to be used through hardware description
families, from low-end 8-bit to high-end 32-bit,
language (HDL) programming and synthesis. It
can be chosen. Another issue to be considered is
has a high unit cost, consumes more power, but
whether the application program is executed in an
it still provides reasonable performance for fast
operating system (OS) or non-OS environment.
67
For a low-end microcontroller, the program is some dedicated hardware is added for specific
usually implemented in a non-OS foreground/ applications, and the performance is improved.
background mode, because the on-chip memory is DSP is also a software approach, because the
small and the system clock is slow, so usually the logic is implemented by software. Developments
control functions are not complicated. On the other on DSP have advantages and drawbacks lying
hand, adopting a high-end 32-bit microcontroller between microprocessor and PLD. Texas Instru-
implies the requirement of faster computations ments’ TMS320 DSP series with a fixed-point or
and complicated functions. Because of larger floating-point provides solutions to automotive
memory and a faster clock, the OS environment, infotainment, vision control, and digital radio.
which takes care of system resource management,
is much better and more convenient for applica-
tion development. sensors
Renesas has several microcontroller families,
for instance: SuperH, H8, and M16/R32 to provide Sensors are applied as the input devices in an
a variety of automotive functions, like powertrain, ECU for obtaining environmental and vehicular
chassis, body, active/passive safety, audio and parameters. The advanced control of a modern
navigation. Since the microcontroller is the core automobile largely depends on the data acquired
of an electronic control unit and connected on the from sensors. Sensors convert a physical parameter
in-vehicle network, CAN or FlexRay, the network to an electrical signal, and a signal conditioning
physical controller is mandatory in the microcon- circuit adjusts the electrical signal to the voltage
troller family for automotive applications. range specified by the analog-to-digital converter
(ADC). After the signal is converted to digital
programmable logic device data, the processor can process it by applying a
simple value comparison or a complicated digital
The system implemented on a programmable signal processing algorithm.
logic device is referred to as a hardware approach, The advancement of electronic technology
because the logic is realized on synthesized logic and material science pushes the innovation of
gates. This approach is usually implemented for sensor development. It improves the functions
prototype designs and verification, and later the and decreases the cost of an automotive system,
logic circuits will be manufactured as a full-custom so vehicle features are greatly enhanced. Sen-
IC. The throughput, power consumption, and cost sors play a more and more important role in
of PLD are usually inferior or more expensive the implementation of ECUs in vehicles. The
than that of a full-custom IC, but it is convenient average number of sensors per vehicle was 40
for prototype development. For example, Xilinx in 2007, and will be up to 70 in 2013, while the
provides XA products with Spartan-3 FPGA and total sensor production will jump from 640 M to
IPs to implement image processing, video, or in- 1100 M in North America. Modern luxury cars
vehicle network solutions. have more than 100 sensors per vehicle, and total
of 167 automotive applications were described.
digital signal processor The applications of sensors in a vehicle can be
categorized into three areas: powertrain, safety,
The digital signal processor (DSP) is an application and comfort (Fleming, 2008).
specific processor, which optimizes the datapath Some sensors used in automobiles are de-
design, so execution of signal processing instruc- scribed more in the following subsections.
tions has a much higher throughput. Moreover,
68
accelerometer are coded and transmitted to central console by

radio frequency (RF) communications (Normann
The accelerometer, or g-cell, is applied for et al., 2003).
crash-detection. A microelectrical mechanical
system (MEMS) accelerometer, which employs Temperature
single-point sensing, reduces the dimension and
cost by removing the wiring and harness (Monk Several temperature sensors are applied in moni-
et al., 2003). toring engine, coolant, gear-box, fuel, intake-air,
Airbag systems, including front airbags, side outside-air, passenger compartment, climate
airbags, seat belt pretensioners and occupant control unit parameter, and exhaust-gas in an
sensing systems, become more complicated and automotive system. There are two temperature
need more sensing requirements. Airbag systems ranges to be measured. The first one covers
are required by government safety regulations, -40 to 170°C, and the second range for exhaust
and the accelerometer plays an important role in temperature measurement covers -40 to 760°C
the operation. (Bojarski, 2003).
yaw-rate radar
Yaw-rate sensor, angular rate sensor, or gyroscope, Radar is employed for collision avoidance by
measure the rotation of a body in angle per unit of finding the distance to the preceding vehicle,
time along a specified axis. Typical applications and ACC can be implemented by utilizing the
include electronic stability program, rollover information of distance and relative speed. Since
protection, and navigation. When concerning, the the vehicle speed is high and response time is
turning movement is monitored and is compared short on highway, a long-range radar (LRR),
with the angle of the steering wheel and vehicle up to 120 m, which takes 3.6 s to collide with
speed. Rollover accidents usually happen during the stationary object at the speed 120 km/hr, is
high-speed driving, and causes serious injury or required for ACC. Short-range radars (SRR), up
death to the driver and passengers, so the yaw-rate to 20 m, are applied for both sides of vehicle to
sensor must recognize the rollover of a vehicle create a safety shield around the car (Knoll, 2003).
and thus activating the safety systems appropriate Figure 10 demonstrates surround sensing using
protection (Schatz et al., 2003). radar and video.
pressure Video
A variety of pressure sensors are employed for Though radar can easily find the distance of an
automotive applications. Vaporized gasoline leak object in front, it has a narrow beam and cannot
detection is in the low-pressure field (Yokomori recognize objects or detect borders. A camera sys-
et al., 2003). Suspension pressure detection and tem is used for several applications in automotive
air-conditioner refrigerant pressure detection system, like lane departure warning, night-vision
are in the high-pressure field. While gasoline improvement, object detection, blind spot warn-
fuel injection is in the very high-pressure field ing, and distance warning.
(Gerbers, 2003). Tire-pressure sensors measure The video solution seems more feasible, but
the air pressure and temperature inside the tire. some issues need to be addressed. The captured
The data along with tire ID and battery lifetime images are essentially affected by light intensity,
69
Figure 10. Surround sensing using radar and video
so an intelligent high-quality camera is required gation system, active front and rear steering, and
for automatic adjustment. A far infrared (FIR) steer-by-wire (Kofink, 2003).
technology is used for night-vision improvement.
Real-time image processing needs intensive com- Torque
puting power, so a high-speed industrial computer,
which may be expensive and occupies large space, The torque sensors measure the powertrain torque
is required (Knoll, 2003). signals on engine and transmission control units,
which are part of the applications of TCS and ESP.
wheel-speed Electric power steering is another application of
torque sensors, which measure angular displacement,
The wheel-speed sensor measures the movement which is proportional to torque (Morbe, 2003).
and circumstance of the tire. The ABS and TCS,
prevents wheel locking and spinning, needs to chemical
know the wheel speed when applying the brake
or accelerator, respectively. ESP for vehicle There are several types of chemical sensors, in-
stability also needs wheel speed information. cluding oxygen detection for air/fuel ratio control
Other applications include transmission control, (Riegel et al., 2003), NOx sensors for emission
odometer, navigation system, stop and go, and control (Schmitt, 2003) and liquid media sensor
roll over protection (Morbe, 2003). (Jakoby et al., 2003). The exhaust gas of a vehicle
has to meet emission limits, so a catalyst system is
steering-angle installed to do the conversion. While oxygen and
NOx sensors are used for after-treatment sensing,
Steering angle sensor is applied for ESP, which liquid sensors are used to detect the quality of
requires steering angle along with yaw rate, wheel liquids, such as engine oil, gear-box oil, fuel, and
speed, and lateral acceleration to determine if battery liquid to check if they should be changed
the vehicle oversteers or understeers. The sensor or are in abnormal states.
is also used in electric power steering, adaptive
cruise control, forward intelligent lighting, navi-
70
acTuaTors ignition
Actuators are output devices to drive electrome- The actuators of the ignition system are the combina-
chanical components. They are either implemented tion of the spark plug, the ignition coil, and driver
by a solenoid, which is controlled by duty cycle electronic circuits. When the controller sends a
or PWM approach, or an electric motor, like the signal turning on the driver, i.e. a power transistor is
stepper motor or dc motor (Bonnick, 2001). conductive, current flows through the primary coil,
Some example applications of actuators are creating a relatively large magnetic field linked to the
fuel injectors, exhaust gas recirculation, ignition, secondary coil. The controller instantly switches off
ABS modulators, variable valve timing (VVT), the signal, causing the transistor to be nonconducting.
and electric motors for hybrid/electric vehicles. The sudden rapid drop in the magnetic field of the
secondary coil generates a very high voltage creating
Fuel injectors the spark across the spark plug electrodes, igniting
the mixture and, finally, initiating the power stroke
Fuel injectors are electrically driven actuators that for the engine (Bonnick, 2001).
regulate the flow of fuel into an engine for engine
control applications. A fuel injector is a solenoid- abs modulators
operated valve, which opens or closes to permit or
block fuel flow to the engine. The valve is attached Anti-lock braking system modulator contains a
to the movable element of the solenoid and is pump driven by an electric motor and various
switched by the solenoid activation. The quantity solenoid-operated valves. Solenoid valves with two
of fuel injected is proportional to the duration when hydraulic connections and two valve positions are
the valve is opened, so the valve can be controlled used. At different degrees of brake slip, the solenoid
by applying a pulse train with a specified duty cycle valves switch to different pressure settings to change
to the solenoid (Bonnick, 2001). the pressure in the brake (Bonnick, 2001).
exhaust gas recirculation

sTorage deVices
Exhaust gas recirculation (EGR) is utilized to
reduce NOx emissions. The engine controller de- Storage devices were not commonly used in
termines the amount of EGR and sends a signal to conventional vehicle systems, because there were
EGR actuator. Typically, this actuator is a variable- few applications involving the storage of large
position solenoid controlled valve that regulates amounts of data. One application for drivers or
the EGR as a function of intake manifold pressure passengers to have some entertaining activities
and exhaust gas pressure. When the EGR valve is during their drive is playing audio or video from
open, exhaust gas flows into the intake manifold. CDs or DVDs. Today, the navigation systems are
Like the fuel injector, by sending a pulse train with more affordable and popular for eliminating the
specified duty cycle to the solenoid, EGR actuator map-reading, so a mass storage device for saving
regulates the pressure to open the valve and then the data of maps and landmark pictures is neces-
the amount of EGR (Bonnick, 2001). sary. More applications for an infotainment system
are developed, while the mass storage devices are
mandatory for storing the multimedia contents.
Hard disks are the most common storage de-
vices in personal computers, but they don’t work
71
in an automotive system, which is operated in and button/knob will be replaced by liquid crystal
an environment with severe vibration. After the display (LCD) and touch screen in the central
solid-state storage technology is more established console. Due to the limited space on dashboards
and the cost is lower, the automotive applications and the increased amount of information that
with mass storage devices are possible. needs to be displayed, the graphic LCD display
Flash memory is based on the solid-state will integrate a variety of information modes into
technology and two types of flash memory, one module and can show the information flexibly.
NAND flash and NOR flash, have emerged as the According to the usage priority or user interac-
dominant varieties of non-volatile semiconduc- tion, the LCD display can show driver functions
tor memories utilized in embedded systems. The and vehicle status, navigation information, and
characteristics of NAND flash are high density, driving assisted images.
medium read speed, high write speed, high erase To simplify the central console appearance,
speed, and an indirect or I/O like access, so it is the LCD display and touch screen panel are inte-
low cost and has been used primarily as a remov- grated into one module. The touch screen panel is
able high-density data storage medium, which is mandatory in mobile devices and eliminates the
appropriate for mass storage applications. The usage of keyboard and mouse in a conventional
characteristics of NOR flash are lower density, PC environment. This intuitively enhances the hu-
high read speed, slow write speed, slow erase man machine interaction on a more user-friendly
speed, and a random access interface, so it has graphical presentation. There are two common
typically been used for code storage and direct types of touch screen: capacitive and resistive.
execution in portable electronics devices. Flash For security applications, biometric system
memory will increase in demand, since it can and radio-frequency ID (RFID) devices are input
store not only multimedia data and navigation modules for identity confirmation and personal-
information, but also the run-time vehicle data ized operation, so they enhance the convenient
and image as an event data recorder. features of automobiles. The biometric mecha-
nisms, including fingerprint identification, face
recognition, iris recognition and voice recognition,
human machine inTerFaces verify if the person’s unique biometric charac-
teristics is matching with the enrollment setting
Most ECUs don’t interact with the users directly, (Robert Bosch, 2006). These approaches have been
so driver and passengers need to control the vehicle researched for a variety of security applications,
through either dashboard or other mechanical but most of them still do not reach a perfect suc-
devices, like the steering wheel, brake pedal, or cessful rate in various environments or don’t have
button/knob. Though the electronic components a straightforward and friendly operation process.
being used are increasing, the conventional pow- Fingerprint identification is a more reliable ap-
ertrain and brake operation will not change in the proach and is employed for personalized adjust-
near future, as a result of driving habits and safety ment functions and configurable favorite settings.
reasons. Steer-by-wire and brake-by-wire haven’t RFID devices, using radio frequency to read and
substituted the steering wheel and brake pedal verify the identification and information, have the
yet. However, the user interface to enhance the same application as the fingerprint approach. This
convenient infotainment management and comfort technology or similar ones will be the innovative
adjustment will be realized and popular. alternative to the car key and be a personal infor-
The human machine interface involves input mation storage device in the future.
and output. The conventional needle instrument
72
reFerences Morbe, M., & Zwiener, G. (2003). Whell-speed

sensors. In J. Marek, H.P. Trah, Y. Suzuki,& I.
Bojarski, A., & Fichte, W. (2003). Temperature Yokomori, (Ed.), Sensors application Vol. 4 -
sensors. In J. Marek, H.P. Trah, Y. Suzuki,& I. Sensors for automotive technology (pp. 403-415).
Yokomori, (Ed.), Sensors Application Vol. 4 - Sen- Berlin: Wiley-VCH.
sors for Automotive Technology (pp. 343-359).
Berlin: Wiley-VCH. Normann, N., Schulze, G., & Keller, W. (2003).
Tire-pressure sensors. In J. Marek, H.P. Trah, Y.
Bonnick, A. (2001). Automotive computer con- Suzuki,& I. Yokomori, (Ed.), Sensors application
trolled systems. London: Butterworth Heine- Vol. 4 - Sensors for automotive technology (pp.
mann. 536-544). Berlin: Wiley-VCH.
Fleming, W. J. (2008). New automotive sensors−A Paret, D. (2007). Multiplexed networks for embed-
review. IEEE Sensors Journal, 8(11), 1900–1921. ded systems – CAN, LIN, FlexRay, Safe-by-Wire.
doi:10.1109/JSEN.2008.2006452 Hoboken, NJ: Wiley.
Gerbers, J. (2003). High-pressure sensors. In Riegel, J., Wiedenmann, H., & Neumann, H.
J. Marek, H.P. Trah, Y. Suzuki,& I. Yokomori, (2003). Chemical sensors for oxygen detection
(Ed.), Sensors Application Vol. 4 - Sensors for and air/fuel ratio contol. In J. Marek, H.P. Trah, Y.
Automotive Technology (pp. 333-342). Berlin: Suzuki,& I. Yokomori, (Ed.), Sensors application
Wiley-VCH. Vol. 4 - Sensors for automotive technology (pp.
Jakoby, B., & Herrmann, F. (2003). Chemical 480-499). Berlin: Wiley-VCH.
sensors for liquid media. In J. Marek, H.P. Trah, Robert Bosch Gmb, H. (2006). Safety, comfort and
Y. Suzuki,& I. Yokomori, (Ed.), Sensors Applica- convenience systems. Hoboken, NJ: Wiley.
tion Vol. 4 - Sensors for Automotive Technology
(pp. 516-526). Berlin: Wiley-VCH. Schatz, O. (2003). Yaw-rate sensors. In J. Marek,
H.P. Trah, Y. Suzuki,& I. Yokomori, (Ed.), Sen-
Knoll, P. (2003). Radar sensors. In J. Marek, H.P. sors application Vol. 4 - Sensors for automotive
Trah, Y. Suzuki,& I. Yokomori, (Ed.), Sensors technology (pp. 297-313). Berlin: Wiley-VCH.
application Vol. 4 - Sensors for automotive tech-
nology (pp. 372-385). Berlin: Wiley-VCH. Schmitt, D. (2003). Chemical sensors for emis-
sion control. In J. Marek, H.P. Trah, Y. Suzuki,&
Knoll, P. (2003). Video sensors. In J. Marek, H.P. I. Yokomori, (Ed.), Sensors application Vol. 4 -
Trah, Y. Suzuki,& I. Yokomori, (Ed.), Sensors Sensors for automotive technology (pp. 500-508).
application Vol. 4 - Sensors for automotive tech- Berlin: Wiley-VCH.
nology (pp. 386-402). Berlin: Wiley-VCH.
Vahid, F., & Givargis, T. (2002). Embedded system
Kofink, P. (2003). Steering-angle sensors. In J. design: A unified hardware/software introduction.
Marek, H.P. Trah, Y. Suzuki,& I. Yokomori, (Ed.), Hoboken, NJ: Wiley.
Sensors application Vol. 4 - Sensors for automotive
technology (pp. 428-449). Berlin: Wiley-VCH. Yokomori, I., & Suzuki, Y. (2003). Pressure sen-
sors. In J. Marek, H.P. Trah, Y. Suzuki,& I. Yoko-
Monk, D., Mladenovic, D., & Skaw, M. (2003). mori, (Ed.), Sensors application Vol. 4 - Sensors
Accelerometers for automotive applications. In for automotive technology (pp. 314-332). Berlin:
J. Marek, H.P. Trah, Y. Suzuki,& I. Yokomori, Wiley-VCH.
(Ed.), Sensors application Vol. 4 - Sensors for
automotive technology (pp. 269-296). Berlin:
Wiley-VCH.
73
74
Chapter 5
Data Communications Inside
Vehicular Environments
Cheng-Min Lin
Nan Kai University of Technology, Taiwan, R.O.C.
Tzong-Jye Liu
absTracT
ZigBee is based on IEEE 802.15.4 which specifies the physical layer and medium access control (MAC)
for low-cost and low-power LR-WPAN. The technology can be applied in intelligent key, A/C opera-
tion and steering wheel inside vehicles. There are two types of devices in ZigBee, FFD and RFD. A
FFD can communicate with RFDs and other FFDs, while a RFD can only communicate with a FFD.
In ZigBee physical layer, it follows IEEE 802.15.4 standard and operates in unlicensed RF worldwide
(2.4GHz global, 915MHz Americas or 868 MHz Europe). A superframe contained an active portion
and an inactive portion is used in the MAC layer of ZigBee. The active portion includes CAP and CFP.
In the inactive partition, the coordinator can enter sleep mode to save its power. Three main topologies
of ZigBee are star, mesh, and tree. However, ZigBee is successfully produced into a low-cost controller
applied for automotive applications, including vehicle control and status monitoring. According to the
forecast of ON World in 2005 (ON WORLD, 2009), the deployed wireless sensing network nodes will
increase to 127 million in 2010 from 1.2 million in 2005. It can be applied in home automation, battlefield
surveillance, health care applications and vehicular environments. A wireless sensor network (WSN)
constitutes a lot of wireless sensing nodes. In addition, a node in WSN consists of one or more sensors,
a radio transceiver, and a microcontroller. The sensor can be used for sensing temperature, pressure,
sound, vibration, motion or position, etc. to collect status from devices or environments. The transceiver
is used to relay the information of the collected status computed by the microcontroller to a center node,
called a gateway or sink. Therefore, a WSN belongs to one type of wireless ad-hoc networks. However,
the nodes in a WSN are usually smaller than that in traditional wireless ad-hoc networks regarding node
size, computing power, memory size, and transmission rage. In other words, the transmission ability,
computing power, and memory size of WSN nodes are limited.
DOI: 10.4018/978-1-60566-840-6.ch005
Data Communications Inside Vehicular Environments
inTroducTion LR-WPAN standard and the ZigBee Alliance

ratified the first ZigBee standard in December
In vehicular environments, a wireless personal area 2004 (Geer, 2005).
network (WPAN) is required because a driver is
a master for the vehicle. The driver interacts with
his/her vehicle according to the collected informa- oVerView oF ieee 802.15.4
tion from WPAN and a controller area network
(CAN). The standard of WPANs is defined in the The IEEE802.15.4 standard is the basis for the
15th working group of the IEEE 802.15. There are ZigBee specification. It specified the physical
six task groups in the working group as shown in layer and MAC layer for LR-WPAN, providing the
the following. fundamental lower layers of WPAN. Upper layers
of the protocol stack include application profiles
• IEEE 802.15.1: Bluetooth defined by ZigBee Alliance. The architecture
• IEEE 802.15.2: Coexistence ZigBee/802.15.4 is shown in Figure 1.
• IEEE 802.15.3: High rate WPANs (HR- The architecture focuses on low-cost, low-
WPAN), UWB speed ubiquitous communication between devices.
• IEEE 802.15.4: Low rate WPANs (LR- Furthermore, the features of IEEE 802.15.4 are
WPAN), ZigBee illustrated as follows according to the mention
• IEEE 802.15.5: Mesh network of LR-WPAN Task Group (The IEEE 802.15.4
• IEEE 802.15.6: Body area network WPAN Task Group, 2009). Its applications are
technologies shown in Figure 2.
In this chapter, we will focus on the introduc- • Data rates of 250 kbps, 40 kbps, and 20
tion of ZigBee based on IEEE 802.15.4, a standard kbps
completed in May 2003 which specifies the physi- • Star or peer-to-peer operation
cal layer and medium access control (MAC) for • Dynamic device addressing
low-cost and low-power LR-WPAN. Although • Two addressing modes are implemented,
ZigBee-style networks created by the Firefly including 16-bit short and 64-bit IEEE
Working Group in 1999 become ZigBee later, the addressing
group does not exist now. Today’s ZigBee was • Support for critical latency devices, such
adopted in 2003 and built on the IEEE 802.15.4 as joysticks
Figure 1. ZigBee/802.15.4 architecture
75
Figure 2. ZigBee applications
• Carrier Sense Multiple Access/Collision Two different types of devices are defined in
Avoidance (CSMA/CA) channel access is an LR-WPAN, a full function device (FFD) and
provided a reduced function device (RFD). A FFD can
• Automatic network is established by the communicate with RFDs and other FFDs, while
coordinator a RFD can only communicate with a FFD. FFD
• Fully handshaked protocol for reliability of can operate in the device, coordinator, and PAN
transmission coordinator modes, while RFD can only operate
• Power management is implemented to en- in the device mode. The comparison between FFD
sure low power consumption and RFD is shown in Figure 3.
• 27 different channels, including 16 chan- The ZigBee Alliance focuses on the network
nels in the 2.4GHz ISM band, 10 channels layer to the application layer. As shown in Figure
in the 915MHz band and one channel in 4, a ZigBee Device Object (ZDO) is a special
the 868MHz band device in a ZigBee network. It is responsible for
a number of tasks including keeping of device
roles, management of requests to join a network,
Figure 3. FFD vs. RFD
76
Figure 4. Outline of the ZigBee stack architecture
device discovery, and security. A ZDO contains Zigbee physical layer

ZDO Management Plane, and also defines the
role of the device in the network such as ZigBee ZigBee is the technology of a novel WSN that
coordinator or end device. A ZDO can discover is designed for the automation and controlling
devices on the network and determine which ap- buildings, consumer electronics, and intra-vehicle
plication services they provide. ZDOs also can wireless communications. The PHY functional-
initialize or respond to binding requests. The last ities of ZigBee are listed below.
function of ZDOs is to establish secure relation-
ship, among network devices. The functionality of • Activation and deactivation of radio
Application Support Sublayer (APS) is to provide transceiver
interfaces between application layer and network • Energy detection within current channel
layer. The services of the interface are offered • Link quality indication for received
by two entities: data service and management. packets
The APS data entity (APSDE) provides the data • Clear channel assessment for CSMA/CA
transmission service, while the APS management • Channel frequency selection
entity (APSME) provides the management service. • Data transmission and reception
The application frame is the environment whose
application objects are hosted in the ZigBee device. ZigBee follows IEEE 802.15.4 standard and
The application objects are responsible for sending operates in unlicensed RF worldwide (2.4GHz
and receiving the data and provide several func- global, 915MHz Americas or 868 MHz Europe).
tions. The first function is to control and manage There are 27 channels allocated in ZigBee stan-
the protocol layers. The second function is the dards as shown in Figure 5. Channel 0 uses the
initialization of standard network functions. The frequency at 868.0 ~ 868.6 MHz, while the data
Security Service Provider is responsible for key rate is 20 kbps. Channels 1 ~ 10 use the frequency
establishment, key transport, frame protection, at 902.0 ~ 928 MHz, where each channel can
and device management. provide 40 kbps data rate. Channels 11 ~ 26 use
the frequency at 2.4 ~ 2.4835 GHz, where each
channel is 250 kbps.
The physical layer of 868/915 MHz and 2.4
GHz uses Direct Sequence Spread Spectrum
77
Figure 5. ZigBee channel overview
Figure 6. PHY frame structure
(DSSS) to carry data to the channels. In 868/915 Zigbee mac layer

physical layer, the packet error rate must be equal
to or smaller than 1% so that the hardware receiver In the ZigBee specification, the superframe format
sensitivity is required to be equal to or higher than is shown in Figure 7. It is defined by the network
-92 dBm. In 2.4 GHz physical layer, hardware coordinator. The superframe duration is the beacon
receiver sensitivity is required to be equal to or interval broadcasted by the network coordinator.
higher than -85 dBm. The transmission power of A superframe contains an active portion and an
ZigBee must be higher than -3 dBm (0.5 mW). inactive portion. The active portion can be divided
During normal operations, the transmission power into 16 time slots, which can be cataloged as con-
is 0 dBm (1 mW). When the transmission power tention access period (CAP) and contention free
is 0 dBm, the transmission range is about 10 ~ 20 period (CFP). The coordinator only receives and
meters. For variant applications, the transmission sends data with other devices in WPAN at active
power can be adjusted so that the transmission portion. In the inactive partition, the coordinator
range can be increased to 30 ~ 75 meters. can enter sleep mode to save its power.
The beacon message is broadcasted at time
The PHY frame structure of ZigBee is shown
slot 0. The purposes of the beacon message are
in Figure 6. There are three fields in the packet.
listed as follows:
The preamble (32 bits) is used for synchroniza-
tion and the start of packet delimiter (8 bits) shall
• Starting superframes
be formatted as “11100101”. The PHY header
• Synchronizing with associated devices
(8 bits) contains the length (0 to 127 bytes) of
• Announcing the existence of a PAN
PSDU data field. • Informing pending data in coordinators
78
Figure 7. The superframe structure
Any device wants to communicate with the a non beacon-enabled network, devices simply
coordinator in the contention access period must transmit their data using unslotted CSMA/CA.
use a slotted CSMA/CA mechanism to access The procedure is shown in Figure 8.
the time slots. For some applications that devices Coordinator to device. In a beacon-enabled
require low delay time or fixed transmission rate, network, the coordinator indicates in the beacon
the coordinator can assign some Guarantee Time that the data is pending. Device periodically listens
Slots (GTS) for them to use. The contention free to the beacon and transmits a MAC command
period consists of several GTSs. The maximum request using slotted CSMA/CA if necessary. In
number of GTS is limited to seven according to a non-beacon-enabled network, a device transmits
the standard. This means that only the limited a MAC command request using unslotted CSMA/
number of nodes can use GTS. CA. If the coordinator has its pending data, the
In ZigBee, there are three data transfer models: coordinator transmits data frame using unslotted
“device to coordinator,” “coordinator to device,”
CSMA/CA. Otherwise, the coordinator would
and “device to device.” The data transfer models
transmit a data frame with zero length payload
are discussed below.
(Figure 9).
Device to coordinator. In a beacon-enable
Device to device. In a peer-to-peer topology,
network, devices would find the beacon to syn-
devices may directly communicate with other
chronize to the superframe structure, and then they
would use slotted CSMA/CA to transmit data. In devices in the transmission range. In order to
transmission data efficiently, the device transmit-
Figure 8. Data transfer model (Device to coordinator)
79
Figure 9. Data transfer model (Coordinator to device)
ting data can not enter the sleep mode. It would • To support star topology and cluster-tree
use unslotted CSMA/CA to transmit data. topology, the routing protocol of ZigBee
uses the concept of tree routing. When a
Zigbee network layer device receives a packet, it first checks
if the device itself or one of its child end
The specification of ZigBee provides three types devices is the destination. If so, this de-
of topologies: star topology, mesh topology, and vice will accept the packet or forward this
cluster tree topology as shown in Figure 10. Ac- packet to the designated child. Otherwise,
cording to the ability of ZigBee devices, we can it would relay the packet along the tree.
divide them into FFDs and RFDs. FFDs have a • To support mesh topology, the ZigBee
lot of resources that include the computing ca- routing protocol uses the concept of Ad-
pability, memory, and the power than RFDs. In hoc On-demand Distance Vector (AODV)
addition, these topologies are made up by three (Perkins et al., 1999), an on-demand ap-
types of devices. First of all, the most important proach for finding routes. Links with lower
type of devices is the ZigBee coordinator (ZC). cost will be chosen into the routing path.
Any topology can have only one ZC. Furthermore, The cost of a link is defined based on the
the ZC in a ZigBee-based network is also a FFD. packet delivery probability on that link.
The ZC is responsible for network formation and Route discovery procedure was discussed
maintenance. The second type of devices is the as follows. The source broadcasts a route
ZigBee router (ZR). The ZR is a FFD or a RFD. request packet. Intermediate nodes will re-
However, the resource in a RFD has less than that broadcast route request if they have rout-
in a FFD. Moreover, the router is responsible for ing discovery table capacities and the cost
forwarding packets in the network. The last type is lower. Otherwise, nodes will relay the
of devices is the ZigBee end device (ZED). The request along the tree. The destination will
ZED is the RFD and it cannot forward packets. In choose the routing path with the lowest
other words, the ZED cannot relay data from other cost and then send a route reply. The ad-
devices and only talk to their parent devices. vantages of AODV are no extra traffic for
80
Figure 10. Three topologies of ZigBee
communication along existing links, and ì

ï
ï1 + C m × (Lm - d - 1), if Rm = 1 (a )
ï
C (d) = ï
í
L -d -1
1 + C m - Rm - C m × Rm m
supporting unicast/ multicast. skip ï
ï , otherwise(b )
ï
ï 1 - Rm
î
When a new ZigBee device joins the network,

it has to choose a new suitable parent node. In
ZigBee specification, the new parent must satisfy Address assignment starts from the ZigBee
three conditions. First, its information is in the coordinator. It would first see its address and
new ZigBee device’s neighbor table. Second, the depth to 0. If a parent node at depth d has an
link cost between the new ZigBee device and the address Aparent the nth child router is assigned
parent node must be more than 3. Finally, the new t o a d d r e s s Aparent + (n - 1) * C (d ) + 1
skip
parent node must have the minimum depth from and n child end device is assigned to address
th
coordinator in the neighbor table. Aparent + Rm * C

skip
(d ) + n . An example of
In ZigBee, network addresses are assigned ZigBee network address assignment was shown
to devices via a distributed address assignment in Figure 11. In this example, the ZigBee coor-
scheme. When the network was formed, the Zig- dinator assigns Cm = 6, Rm = 4, and Lm = 3. The
Bee coordinator would determine three network Cskip value can be calculated by above equation
parameters. and the result is 31. The addresses of the first to
third child router of the coordinator were 0 + (1-
• Cm: Maximum number of children of a 1)31+1=1, 0 + (2-1)31 + 1 = 32 and 0 + (3-1)31
ZigBee router + 1 = 63. The addresses of the two devices of the
• Rm: Maximum number of child routers of coordinator were 0 + 431 + 1 = 125 and 0 + 431
a parent node + 2 = 126.
• Lm: Depth of the network
The ZigBee Cluster Label
ZigBee defines Cm“≥ Rm, so that at least Cm“-Rm
ZigBee devices can connect to a ZigBee router. A K.K. Lee et al. used ZiCL (Lee et al., 2006) to
parent device utilizes Cm, Rm, and Lm to compute a improve the shortcoming of the AODV (Perkins
parameter called Cskip, which is used to compute et al., 1999) that has higher routing overhead pro-
the size of its children’s address pools. The Cskip duced in route discovery phase. ZiCL divides the
at depth d is computed as follows. ZigBee topology into several logical clusters. A
cluster consists of a clusterhead and some cluster
members. A clusterhead represents a cluster and
81
Figure 11. An example for address assignment
cluster members are grouped into a cluster with 12, node Q will send data to node J that node Q
a clusterhead. Each cluster has a unique cluster is the source node and node J is the destination
label that is assigned to each clusterhead. Some node. First, node Q sends a RREQ packet with
rules must be confirmed for forming a cluster: the destination address of node J. The neighbor
nodes of node Q receive an RREQ packet and
• Only the coordinator or routers can gener- rebroadcast it to the network until the RREQ
ate a logic cluster. packet reaches to node J. When intermediate
• The coordinator must be a clusterhead. nodes receive the RREQ packet, they would add
• Routers with even depth value can be an entry for the Cluster Label 0x0002 in their RT
clusterheads. and add an entry for destination address 0x002C
• Routers with odd depth value would join in their RDT. After node J receives the RREQ
their parent’s cluster. packet, node J would forward the RREP packet
• End devices would join their parent’s along with the reverse path created by the RREQ
cluster. packet. The intermediate node H does not add the
Cluster Label in its RT because the address of the
To follow the rules, the network will form sev- node H is equal to the destination node’s Cluster
eral clusters. Figure 12 shows a Cluster Labeled Label 0x0002. Node H can send data packets
network. Each router node has a routing table (RT) through the intra-routing. When node B and node
and a route discovery table (RDT). The address A receive the RREP packet, they make their rout-
information of Cluster Label is stored in RT and ing entries active and forward the RREP packet to
route discovery entries are stored in RDT for the node F. Node F receives the RREP packet and it
original destination node address. A 16-bits ad- has information of adjacent node Q. Node F sends
dress in ZigBee network is assigned by the ZigBee the RREP to node Q. Finally, node Q establishes
specification (Zigbee Alliance, 2006). In Figure a route for Cluster Label 0x0002. At the same
82
Figure 12. The ZiCL network
time, node Q sends a RNOT packet to notify its following the parent node, and goes down to the
clusterhead G about the Cluster Label 0x0002 destination node. The data packet needs 4 hops
information. Node G gets the RNOT packet then to reach the destination node. Taehong Kim et
it uses the Cluster Label Broadcasting to broadcast al. proposed the shortcut tree routing (Kim et al.,
the RUPT packet to share the routing information 2007) to improve the shortcomings of the ZigBee
with node T and node S. Node T and node S can tree routing. Figure 13 (b) shows the concept of
maintain the routing entry for the Cluster Label the shortcut tree routing, where the source node
0x0002. If node T or node S has child nodes that can send data packets directly to the destination
want to send data to the destination node J, the node. In this way, just one hop is required to reach
routing path will not be established. The child the destination node. The shortcut tree routing can
nodes of the node T or node S can use the cluster overcome the routing overhead of the tree routing
label information in the RT. The nodes can save algorithm. The shortcut tree routing algorithm
power and send data packet faster. basically follows the ZigBee tree routing algo-
rithm with the neighbor table which is defined
Shortcut Tree Routing in in ZigBee specification.
ZigBee Networks The shortcut tree routing algorithm first
chooses a node from the neighbor table to be the
In the cluster-tree topology of the ZigBee network, next hop node. It computes the remaining hop
the routing path used by a tree-based routing pro- counts from the next node to the destination for
tocol may be too long. If the destination node is all the neighbor nodes, including parent and child
near the source node in the cluster-tree network, nodes. In the neighbor table, it would choose the
the source node must pass packets to its node next hop node that can reduce the routing cost.
parent to the destination node. Figure 13 (a) is In Figure 13 (b), it shows that all of the neighbor
an example of the tree routing. The source node nodes of the remaining hops to destination node
sends the data packet to destination. The data were computed. In Figure 13 (b), the source node
packet from the source node goes up to root node transmits data packets to the destination node di-
83
Figure 13. Routing protocols
rectly if the route cost is minimized. The ZigBee should stimulate the market’s growth when the
tree routing always transmits data packets to the market matures. Hence, these technologies, such
parent node. However, the shortcut tree routing as ZigBee, are offering automakers and their sup-
can be used to find a low cost route to save power pliers many possibilities to enhance the potential
and efficiently transmit data packets. of their products in the future.
Next, we will focus on the technology of
ZigBee for in-vehicular environments. Tsai et
applicaTions oF Zigbee al. (2007) reported the results of a ZigBee-based
case study conducted in a vehicle. According to
IEEE 802.15 standards have defined three pro- their results of the experiments and measurements,
tocols of low cost wireless communications, ZigBee was illustrated to be a viable and promising
including Bluetooth (802.15.1), UWB (802.15.3), technology for implementing an intra-car wireless
and ZigBee (802.15.4) that can be used to per- sensor network. According to Table 5.2, Bluetooth
form automotive functions inside a vehicle. The for voice transmission is better than ZigBee due
comparisons among them are listed in Table 1. to a lager link bandwidth provided. However,
These wireless technologies can be applied in the simulation results proposed by Wang et al.
entertainment devices, handsets, intelligent key, demonstrate that ZigBee can support a limited
A/C operation and steering wheel inside vehicles. range of voice services (Wang et al., 2008). From
Bickel surveyed state-of-the-art technology in their experimental results, two directly connected
wireless communication technology within ve- ZigBee nodes can support up to three voice over
hicles, as well as between vehicles (Bickel et al., IP (VoIP) and 17 half-duplex push-to-talk (PTT)
2006), and pointed out that although information, conversations.
telematics, and mobile commerce are the most
prominent application areas for Bluetooth, WiFi,
and WiMax, the fact that there are many differ-
ent technologies of wireless networks in vehicles
84
Table 1. Comparison among Bluetooth, ZigBee, and UWB (Akingbehin et al, 2005)
Bluetooth ZigBee UWB

Range 10m 10m <10m
Chip Price $5 $2 $1
Data Rate Medium Low High
Throughput Medium Low High
Interference Good Good Excellent
Media Voice/Data Data Video/Radar
SIG Consortium Alliance Forum
Data Payload 2744 104 Evolving
TxPower 1 mW < 1 mW 200 µW
Mode FHSS DSSS DS, MBOA
Frequency 2.4 GHz .8, .9, 2.4 G 3.1-10.6 G
Channels 23 or 79 1, 10, or 16 Evolving
Error Correct 8-bit, 16-bit 16 CRC Evolving
Security Good Good Excellent
Topology Star Star, Mesh, Tree Peer-to-Peer
# of nodes 7, or more 65534 Evolving
Link BW 1 MHz 20-250 KHz 120M-1GHz
implemenTaTion oF Zigbee implementation needs a flash memory from 64K

to 96K bytes. Based on the requirement of appli-
A transceiver is a combination of transmitter and cations, there are several special functions, such
receiver in a single package. In a radio transceiver, as flash memory size, ADC support, GPIO and
the receiver is silenced while transmitting. An operating voltage, need to be considered.
electronic switch allows the transmitter and re- The implemented platform is an autonomous,
ceiver to be connected to the same antenna, and
prevents the transmitter output from damaging
of the receiver. ZigBee chip vendors typically Figure 14. ZigBee modules
sell integrated radios and microcontrollers with
60K ~ 128K bytes flash memory, such as the
Maxstream Xbee, the MeshNetics ZigBit and the
CEL ZFSM-100 as shown in Figure 14. Radios
were also used stand-alone with any processors
or microcontrollers. Generally, the chip vendors
also offer the ZigBee software stack, although
independent ones are also available.
The RF transceiver is just one component in
a ZigBee platform solution. A processing device,
such as an MCU, is required to complete the entire
solution by housing the IEEE 802.15.4 MAC and
ZigBee software. A typical ZigBee protocol stack
85
Figure 15. The implemented ZigBee platform block diagram
lightweight wireless communication and com- high performance 8 bit RISC microcontroller with
puting platform based on MaxStream Inc. Xbee 64K bytes The peripheral of Atmega64L contains
radio module and a microprocessor. The platform two 8 bit timers, 8-channel 10-bit ADC, dual pro-
has no integrated sensors, since individual sen- grammable serial USARTs and Master/Slave SPI
sor configurations are required depending on serial interface.
the application. Instead, through predetermined For the IEEE 802.15.4 transceiver chip selec-
connector, the platform can be used with various tion, we used a MaxStreem incorporated product,
serial devices, such as digital/analog sensors and the Xbee OEM RF module. The module of Xbee
SPI compatible devices. supports two operation modes:Transparent mode
Figure 15 shows an implemented ZigBee system and Application Programming Interface (API)
block diagram, where Ateml Atmega64L is used mode. The Xbee module operates in transparent
as the system microcontroller. The Atmega64L is a mode defines a set of AT command for controlling
Figure 16. ZigBee platform implementation
86
Table 2. Components of implemented platform
Part Type Designator Footprint Part Type Designator Footprint Part Type Designator Footprint
0.1uF C13 402 4.7uF/16V C14 0805C 6.8uH L2 C3-Y1.5R
0.1uF/16V C17 402 10uF/10V C15 0805C DB9 CON1 DB-9/M
0.1uF/16V C20 402 10uF/10V C16 0805C DCJACK J5 DCJACK
22pF/50V C23 402 104p C2 0805C 1N4148 D1 DS
400K/1% R10 402 10uF/10V C21 0805C AVR_ISP Jtag1 FKV10SN
10K R8 402 10uF/10V C22 0805C CON10 J1 IDC-10
100K/1% R9 402 104p C3 0805C CON10 J2 IDC-10
0.1uF/25V C10 0603C 0.1u/16V C4 0805C JUNPER J4 IDC-3
0.1uF/25V C11 0603C 10uF/10V C5 0805C CON6 J3 IDC-6
0.1uF/25V C18 0603C 20P C6 0805C 10uH L1 L2520
1uF/10V C19 0603C 20P C7 0805C JUMPER JP2 sip2
0.1uF/25V C8 0603C GREEN LED1 0805LED JUMPER JP3 sip2
0.1uF/25V C9 0603C GREEN LED2 0805LED MAX3232 U3 SO16
27R R11 0603R RED LED3 0805LED RT8008 U2 SOT-25
27R R12 0603R 47 R1 0805R ATMEGA64L U1 TQFP-64
0R R6 0603R 47 R2 0805R 8.00MHz XTAL1 X1
0R R7 0603R 47 R3 0805R XBEE U4 XBEE
0.1u/16V C1 0805C 10K R4 0805R
10uF/10V C12 0805C 4.7K R5 0805R
the internal ZigBee protocol stack. Via the AT com- reserved for changing other IEEE 802.15.4 radio
mand, the user can quickly establish an application transceivers, such as AT86RF230.
without the detailed knowledge of ZigBee protocol Table 2 lists the components of the imple-
stack. In the API mode, the Xbee module acts as mented platform. Figure 17 is the platform circuit
a pure IEEE 802.15.4 RF module. Only a serial diagram which was layout by Protel 99SE. The
communication protocol is defined in the API connection between MCU and XBee is via the
mode. Hence, the full ZigBee stack is required to UART1 of MCU. Regarding the power source,
be implemented in applications and the protocol is there are two sources (main and battery power)
used to communicate with the Xbee module. that can be switched via jumper 4.
In Figure 16, it shows a ZigBee platform
implementation. The implemented system has two
power sources: mains-power and battery-power. summary
The power source can be selected from jumper 3.
LED3 is a power indicator. LED2 and LED1 are The chapter introduced the standard of low rate
connected to GPIO port C of Atmega64L. These WPANs to be called ZigBee. Two types of devices,
two LEDs can be utilized for general purpose. FFD and RFD, are discussed. We have known that a
The system reserves a 8-bit GPIO port and a 8-bit FFD can communicate with RFDs and other FFDs,
ADC connector for extensions. The SPI is also while a RFD can only communicate with a FFD.
87
Figure 17. Circuit diagram of the implemented platform
In ZigBee physical layer, ZigBee follows IEEE reFerences

802.15.4 standard and operates in unlicensed RF
worldwide (2.4GHz global, 915MHz Americas Akingbehin, K. (2005). Wireless communications
or 868 MHz Europe). A superframe contained for intra-vehicle use (Wireless harnesses). Insti-
an active portion and an inactive portion is used tute for Advanced Vehicle Systems, University of
in the MAC layer of ZigBee. The active portion Michigan-Dearborn, Dearborn, MI.
includes CAP and CFP. In the inactive partition, Bickel, G. S. (2006). Inter/intra-vehicle wireless
the coordinator can enter sleep mode to save its communication. Retrieved from http://userfs.cec.
power. Three main topologies of ZigBee are star, wustl.edu/~gsb1/index.html
mesh, and tree. However, ZigBee is successfully
produced into a low-cost controller applied for Geer, D. (2005). Users Make a Beeline for ZigBee
automotive applications, including vehicle control Sensor Technology. Computer, 16–19.
and status monitoring.
Kim, T., Kim, D., Park, N., Yoo, S., & Lopez, T. S.
(2007). Shortcut tree routing in ZigBee networks.
In . Proceedings of ISWPC, 07, 42–47.
88
Lee, K. K., Kim, S. H., Choi, Y. S., & Park, H. Perkins, C. E., & Royer, E. M. (1999). Ad-hoc
S. (2006). A mesh routing protocol using cluster On-demand distance vector routing. In . Proceed-
label in the ZigBee network. In Proceedings of ings of WMCSA, 99, 90–100.
Mobile Ad-hoc and Sensor Systems (MASS), (pp.
The, I. E. E. E. 802.15.4 WPAN Group. (2009).
801-806).
Retrieved from http://www.ieee802.org/15/pub/
TG4.html
89
90
Chapter 6
Wireless Access in
Vehicular Environments
Tzong-Jye Liu
Ching-Wen Chen
absTracT
The IEEE 1609 standards define communication for wireless access in vehicular environment (WAVE)
services, which enable vehicle-to-vehicle, vehicle-to-roadside, as well as vehicle-to-infrastructure com-
munications. The standard consists of four parts, which are briefly described in this chapter. IEEE 1609.1
describes the WAVE resource manager which specifies the wireless access method in a WAVE environment
and allows a remote manager application to establish connection with a resource command processor on
an on-board unit. IEEE 1609.2 defines several secure message formats to process messages for WAVE
system. The standard covers methods for securing WAVE management messages and application mes-
sages, which protects messages from attacks such as eavesdropping, spoofing, alteration, replay, and
linkable information to unauthorized parties. IEEE 1609.3 defines network services for WAVE systems,
whose network services operate at the network and transport layers of the OSI model and support both
the IPv6 traffics and the WAVE short message services. IEEE 1609.4 describes WAVE multi-channel
operations. It specifies the functions of MAC sublayer management entity and WAVE MAC with channel
coordination. The multi-channel operation provides an efficient mechanism that controls the operation
of upper layer across multiple channels.
inTroducTion wireless access in vehicular environment (WAVE)

services, which enable vehicle-to-vehicle, vehicle-
The IEEE 1609 standards (IEEE 1609.1TM, 2006; to-roadside, as well as vehicle-to-infrastructure
IEEE 1609.2TM, 2006; IEEE 1609.3TM, 2007; communications. The standard consists of four
IEEE 1609.4TM, 2006) define communication for parts, which are briefly described in this chapter.
IEEE 1609.1 (IEEE 1609.1TM, 2006) describes the
DOI: 10.4018/978-1-60566-840-6.ch006 WAVE resource manager which specifies the wire-
Wireless Access in Vehicular Environments
less access method in a WAVE environment and messages for DSRC/WAVE system are defined.
allows a remote manager application to establish Section 6.5 describes the IEEE 1609.1 (2006), the
connection with a resource command proces- WAVE resource management is introduced.
sor on an on-board unit. IEEE 1609.2 (IEEE
1609.2TM, 2006) defines several secure message
formats to process messages for WAVE system. waVe mulTi-channel
The standard covers methods for securing WAVE operaTions
management messages and application messages,
which protects messages from attacks such as IEEE 1609.4 (2006) describes WAVE multi-
eavesdropping, spoofing, alteration, replay, and channel operations. It specifies the functions of
linkable information to unauthorized parties. IEEE MAC sublayer management entity (MLME) and
1609.3 (IEEE 1609.3TM, 2007) defines network WAVE MAC with channel coordination. Multi-
services for WAVE systems, whose network channel operation provides an efficient mechanism
services operate at the network and transport lay- that controls the operation of upper layer across
ers of the OSI model and support both the IPv6 multiple channels. The channel coordination
traffics and the WAVE short message services. enhances the mechanism defined in the MAC of
IEEE 1609.4 (IEEE 1609.4TM, 2006) describes IEEE 802.11 and interacts with IEEE 802.2 logical
WAVE multi-channel operations. It specifies the link control and IEEE 802.11p PHY.
functions of MAC sublayer management entity WAVE devices (on-board units or roadside
and WAVE MAC with channel coordination. The units) provide an architecture that supports a single
multi-channel operation provides an efficient control channel (CCH) and multiple service chan-
mechanism that controls the operation of upper nels (SCHs). The control channel is for transmit-
layer across multiple channels. ting WAVE short message and announcing WAVE
This chapter describes the standards for the services. The service channels are for interactions
wireless access in vehicular environment (WAVE). and transmissions between applications. WAVE
The IEEE 1609 standards define communica- standard uses the specification of the PHY in IEEE
tion for WAVE services, which enable vehicle to 802.11 and revises to IEEE 802.11p.
vehicle (V2V), vehicle to roadside, and vehicle The services defined in IEEE 1609.4 (2006) are
to infrastructure (V2I) communication. The com- for managing the channel coordination and sup-
munication system integrates the information of porting MAC service data unit (MSDU) delivery.
engine, gearing, brake, roadside unit, and provides The services include the channel routing, user
safety services for drivers. In the standard, the priority, channel coordination and MAC service
specification of the physical layer is defined in data unit transfer. In the following of this section,
IEEE 802.11p and the communication protocol for we will describe these four services.
WAVE network service is IPv6. This chapter also
describes the four parts of the IEEE 1609 standards. The channel routing
Section 6.2 describes IEEE 1609.4 (2006), the
functions of MAC sublayer management entity WAVE supports both the WAVE short message
(MLME) and WAVE MAC with channel coordi- (WSM) and IP datagram transfer. When an MSDU
nation are introduced. Section 6.3 describes the is passed from the LLC to the MAC, the MAC
IEEE 1609.3 (2007), the WAVE network services determines whether the MSDU is a WSM or an
for WAVE systems are introduced. Section 6.4 IP datagram by checking the EtherType field in
describes the IEEE 1609.2 (2006), the secure the MSDU. If the value of this field is 0x86DD,
message formats and the process of the secure an IPv6 header follows. If the value is 0x88DC,
91
a WAVE short message protocol (WSMP) header data buffer that corresponds to the current service
follows. In the following of this subsection, the channel. If there is no transmitter profile registered
channel routing for WSMP data and IP datagram or the transmitting WAVE device is not a member
is described. of any WAVE Basic Service Set (discussed in
6.3.2), the datagram is discarded.
Channel Routing for WSMP Data
user priority
As is shown in Figure 1, WSMP header contains
the channel, power level and data rate associated IEEE 802.11e Enhanced Distributed Channel Ac-
with the data packet. cess (EDCA) mechanism is used to contend for
The scenario of channel routing for WSMP accessing medium. The MAC buffers the data by
data is as follows: First, WSMP data is passed mapping its user priority to access category index
from the LLC to the MAC. Then, MAC routes (ACI) is defined in IEEE 802.11.
the packet to a proper queue according to the The general architecture of prioritized access
channel number contained in the WSMP header. on one channel for the data transmission is shown
If the channel number does not corresponding to in Figure 2.
the control channel number or the current service When MAC receives a MSDU and completes
channel number, it is invalid. The data packet is the channel routing process, MAC maps its user
discarded if the channel number is invalid. priority to access category index. Each access
category has a unique and independent channel
Routing for IP Datagram function. The channel function is used to pick a
data packet from the access category to compete
The transmitter profile must be registered to the access right. Then, the package being picked
MAC sublayer management entity (MLME) be-
fore initializing IP datagram exchanges. The trans-
Figure 2. Prioritized access for data transmission
mitter profile contains a service channel number,
on one channel. (IEEE 1609.4TM, 2006)
power level, data rate and the adaptable status of
power level and data rate. At any given time, only
one transmitter profile may be active.
Then, IP datagram is passed from the LLC to
the MAC; and the MAC routes the packet to a
Figure 1. Portion of WSMP header used to control

PHY (IEEE 1609.4TM, 2006)
92
Figure 3. Sync interval, guard interval, CCH interval, SCH interval (IEEE 1609.4TM, 2006)
decides which packet gets the access right accord- • IP or WSM requests for data transmission.
ing to its back-off. The packet with the smallest • LLC passes an MSDU to MAC; and the
back-off gets the access right. channel router in MAC distinguishes
between IP and WSMP by checking
channel coordination EtherType field.
◦ If the value of EtherType is WSMP,
Now, we would describe the channel coordina- the channel router assigns the MSDU
tion function, which is done mainly according to to the access category (AC) based on
the MAC layer synchronizing operation. Based the channel number and the user pri-
on the synchronization operation, packets could ority in the WSMP header.
be sent from MAC layer into wireless channel. ◦ If the value of EtherType is IP, the
Shown in Figure 3 is the sync interval, it contains channel router allocates the service
CCH interval and SCH interval components. A channel, access category and user pri-
buffering time interval, called Guard Interval, is ority to the MSDU. Then, the channel
used to synchronize various devices (for example, router puts the MSDU to the channel
synchronization of their time). buffer according to its user priority.
For WAVE devices, Coordinated Universal ◦ When the data unit wins the conten-
Time (UTC) is used as the reference time. The tion on the current channel,
length of the sync interval is dot4SyncInterval; ◦ If the data unit is WSM, the packet
and sync interval starting at the instant the time is transmitted by using the transmit
synchronization function timer modulo dot4Syn- power and the data rate recorded in
cInterval is zero. All WAVE devices shall monitor the header.
the CCH during the CCH interval. ◦ If the data unit is IP datagram, the
packet is transmitted by using the
mac service data unit Transfer transmit power and the data rate
stored in the registered transmitter
In this subsection, we describe the MAC service profile.
data unit transfer defined in IEEE 1609.4 (2006). • The value of the power level and the data
As it was shown in Figure 4, the data transmission rate is set in TXVECTOR. TXVECTOR is
processing flow is as follows: a set of parameters that the MAC provides
93
Figure 4. Data transmission processing flow (IEEE 1609.4TM, 2006)
PHY to transmit the data unit. A clear chan- establishes communications with the provider on
nel is reported by sending an IDLE com- the specified service channel.
mand from PHY to MAC.
• MAC sends a TXVECTOR command for
setting; and a confirmation is replied after waVe neTwork serVices
PHY completes the setting. PHY com-
pletes the setting according to the power IEEE 1609.3 (2007) defines network services for
level and the data rate. WAVE systems. WAVE network services operate
• Data is exchanged between PHY and MAC at the network and transport layers of the OSI
through a series of actions. model, supporting a high data rate, low latency
• A confirmation is replied after PHY com- communication between WAVE devices. The
pletes the transmission. WAVE network services support both the IPv6
• At the end, LLC receives an indication traffics and the WAVE short message (WSM)
from MAC. services.
Now, consider the data transfer in the control overview

channel. WSMs conforming to the WSMP can
be exchanged directly; and the transmission of IP The WAVE network services provide addressing
datagram is not permitted. Now, let us consider and routing services in a WAVE system. It operates
the data transfer in the service channel. In order at the network and transport layers and supports
to initiate communications on a service channel, a wireless connectivity among vehicle-based de-
roadside unit or an on-board unit transmits WAVE vices. It also supports connectivity between fixed
announcement action frames on the control chan- roadside devices and vehicle-based devices. The
nel to advertise offered services available on that wireless communication is based on the 5.9 GHz
service channel. An on-board unit receives the an- DSRC/WAVE mode.
nouncement on the control channel and generally
94
Figure 5. WAVE protocol stack (IEEE 1609.3TM, 2007)
The scope of IEEE 1609.3 (2007) standard is short message protocol has to support the
shown in Figure 5. The WAVE system supports forwarding function.
both IP and non-IP applications. The communi- • Management plane: The management
cations for non-IP based applications are based plane of the WAVE network services
on the WAVE short message protocol (WSMP) provides the following services: applica-
and the IP based applications are based on IPv6. tion registration, WAVE Basic Service
In IEEE 1609.3 (2007), the standard defines the Set (WBSS) management, channel usage
management information base (MIB) of the WAVE monitoring, IPv6 configuration, Received
management entity (WME). It also specifies the Channel Power Indicator (RCPI) monitor-
functions of LLC, UDP/TCP and WAVE short ing, and the Management Information Base
message protocol. (MIB) maintenance.
The WAVE protocol stack consists of data
plane and management plane. About the communication protocol of the
WAVE system, we can go through the channel type
• Data plane: The data plane consists of and two WAVE-supported protocols for explana-
the communication protocol and hard- tion. The WAVE-supported protocol contains both
wares for data transfer. The data plane of the WAVE short message protocol and standard
the WAVE network services must support IPv6 protocol.
UDP protocol; and the TCP protocol is op-
timal. The WAVE network services shall • Channel types: The WAVE system sup-
support the IPv6 protocol and the Logical ports two types of channels: the control
Link Control (LLC) protocol as speci- channel (CCH) and the service channel
fied in IEEE 802.2. The data plane of the (SCH). In a WAVE system, there is only
WAVE network services also supports the one control channel and multiple service
WAVE short message. The WAVE short channels.The control channel is reserved
message can be delivered to multiple desti- for system control message and short, but
nations. The implementation of the WAVE high-priority applications. The service
95
channels support general purpose data nel and distributes this information by using the
transfer applications. control channel.
• WAVE short message protocol (WSMP): WBSS used role-based for controlling regard-
The WAVE short message protocol is de- less of users’ identity, restricting users’ access
signed for optimized operation in the rights. Applications could integrate in the priority
WAVE environment. The WAVE short designing, for example if two applications request-
message may be sent on both the control ing for the accessing of the same channel, WAVE
channel and the service channels. In a management entity would give the access accord-
WAVE system, the WAVE short message ing to the application’s access level. The access
is used for message transmission. It also al- rights of lower layer are decided and controlled
lows application to directly control physi- by MAC layer.
cal layer parameters, such as the channel Data transfer by an application can be operated
number and/or the transmitter power. It is both with a WBSS and without a WBSS.
designed to minimize to channel capacity.
Based on the Provider Service Identifier • Operation without a WBSS: Operation
(PSID), the WAVE short message may be without a WBSS uses WSMP to exchange
sent to the correct destination. data on the control channel. A source appli-
• Standard Internet Protocol (IPv6): The cation composes the WAVE short message
transmission based on the standard IPv6 data and addresses it to the broadcast MAC
can be only used on the service channels address. Then, the application selects ap-
(SCH). The protocol is for generic applica- propriate radio channel information to
tions and network services. control the transmission, and requests the
WSMP for delivery the message. A receiv-
wave system operations ing device accepts the packet and passes
it up to the protocol stack. The WSMP
In a WAVE system, a WAVE Basic Service Set protocol stack delivers it to the registered
(WBSS) is established to support data traffics application(s), based on PSID. Then, the
between applications. A set of cooperating WAVE receiving application knows the address
stations consists of a single WBSS provider and of the provider device, and may exchange
none or multiple WBSS users. Applications based data on the CCH if desired.
on the WAVE short message protocol may initiate • Operation with a WBSS: There are
a WBSS to allocate a service channel, but this is two types of WBSS: persistent and non-
not required since the WAVE short messages may persistent. A persistent WBSS may sup-
be exchanged on the control channel. port Internet access; and a non-persistent
We would move on to the communication WBSS may support on-demand services.
concept of WAVE. Applications themselves can A persistent WBSS is announced during
decide whether WBSS would be employed. When CCH interval and the usage of the persis-
WBSS is not used, WSMs could use only CCH. tent WBSS will offer an ongoing service
If an application exchanges data with a WBSS, to any devices that come into the range
it may send the information by using WAVE of WBSS. Non-persistent WBSS is an-
short message protocol or IPv6 protocol on the nounced only on initiation. The usage of
service channel. A WAVE device may announce the non-persistent WBSS will support an
a WAVE Basic Service Set on the service chan- on-demand service.
96
WAVE applications offer services to potential as eavesdropping, spoofing, alteration, replay, and
user applications by announcing their air interface linkable information to unauthorized parties.
via a WAVE service information element (WSIE). The standard specifies the secure message
WAVE service information element is inside a format by using a presentation language based
WAVE announcement frame. The process of creat- on TLS (IETF RFC 2246). The statement of the
ing a WSIE and transmitting it is initiated when an presentation language includes variable names,
application request to initiate a WBSS and offer data types and functions. Variable names are all
services. On receiving of the WME application lowercase. Multiple words name is indicated by
request, the WME starts the WBSS. underscores. Data types begin with an uppercase
Upon receiving of the notification of the WBSS letter; they may contain a mixture of upper and
initiation from the WME, an application is free lowercase.
to generate data packets (WSM or IPv6 packet)
for transmission on the service channel. Any re- secured messages
ceived packets destined for the application will
be delivered to the application via the WSMP or Many of the applications discussed in this standard
IPv6 stack. The WBSS remains active at the local use the secured message format in Figure 6. The
device until it is ended. generic message format includes the protocol
The WME would terminate a WBSS if any of version and the type. The protocol version is the
the following reason holds. current version of the protocol. The type contains
the type of the message; it tells the receiver how
• All applications have completed their ac- to interrupt the received message. The type value
tivities on the WBSS. The WBSS is no 0 denotes unsecured, the type value 1 denotes
longer being used by locally-registered signed, and the type value 2 denotes encrypted.
applications. Type values from 240 to 255 are reserved for
• A higher priority application participates private usage.
the WBSS and induces the conflict. The message types are as follows (IEEE
• If the lower layer indicates that the service 1609.2TM, 2006):
channel is idle, a user device may termi-
nate its participation in the WBSS. • SignedMessage, ToBeSignedMessage,
and MessageFlags types: These types are
waVe securiTy serVices used when the type field is signed.
• S i g n e d W S M & To B e S i g n e d W S M
IEEE establishes IEEE 1556 to develop the stan- types: The ToBeSignedWSM is the vari-
dard for WAVE security service. Later, the standard ant of the ToBeSignedMessage except
was renamed as IEEE 1609.2 (2006). The main that the application field is omitted. The
goal of this standard is to define the secure message SignedWSM types is also the variant of
formats to process messages for DSRC/WAVE the SignedMessage, expect that the signed
system. The standard covers methods for securing data is a ToBeSignedWSM structure.
WAVE management messages and application • PublicKey, PKAlgorithm, & SymmAlg
messages. The exception of vehicle-originating orithm types: The PublicKey structure is
safety messages and the administrative functions for encoding a public key and identifies the
for the core security functions are also described. algorithm that public key is used.
The standard protects messages from attacks such • ECPublicKey type: An ECDSA or ECIES
public key is specified.
97
Figure 6. Secured message format (IEEE 1609.2TM, 2006)
• CertID8 & CertID10 types: These types • EncryptedMessage, EncryptedCont

are used to identify a certificate. The entInfo, & RecipientInfo types: The
CertIDx, x is 8 or 10, for a given certificate EncryptedMessage type is supported to en-
is calculated by the SHA-256 hash func- crypt a message to one or more recipients
tion and the low-order x bytes of the hash using the recipients’ public keys.
value is output. • ECIESNISTp256EncryptedKey &
• The ApplicationID & FullySpecifiedAp AESCCMCiphertext types: The
pID types: The security services use these ECIESNISTp256EncryptedKey type is
types to determine whether a message may used to transfer the symmetric key. The
legitimately address a given application. symmetric key is encrypted using ECIES
• Time64 & Time32 types: Time64 is a as specified in IEEE Std 1363a-2004.
64-bit integer. It is encoded in big-endian • WAVECertificate, ToBeSignedWAVECe
format and gives the number of microsec- rtificate, CertSpecificData, SubjectType,
onds. Time32 is a 32-bit integer. It is also & CRLSeries types: The standard sup-
encoded in big-endian format and gives the ports many certificate types. The certificate
number of seconds. type corresponding to different roles that
• SignerInfo type: The receiver uses signers can play in the system. All these
SignerInfo structure to determine which certificates use the same structure.
keying material is used to authenticate the • WAVECRL,ToBeSignedCRL, CRLType,
message. and IDAndDate types: These types are
• Signature type: The actual signature is used for certificate revocation list.
contained in the signature field. • WAVE Certi fi ca teReq u es t & WAV
• ECDSASignature type: The ECDSA sig- ECertificateResponse types: The
natures are performed. WAVECertificateRequest type is used
to request a WAVECertificate. The
98
WAVECertificateResponse is used to re- The steps to process the signed message are
turn a WAVECertificate. as follows:
• GeographicRegion & RegionType types:
These types are for defining region. 1. Decode the received octet string.
• The 2DLocation & 3DlocationAndConfi 2. If the type of the application field in the
dence types: 2Dlocation is used to define signed message is not correct, discard the
validity regions using in the certificates. message.
3DlocationAndConfidence is used to iden- 3. Check and ensure that the message is not a
tify the transmission locations and includes replay message.
a confidence field. 4. If the transmission location is included in the
• Certificate scopes: This type defines the signed message, perform the geographical
scope of the certificate. validity checks.
5. Use the signer digest to retrieve the certificate
signed messages from the message certificate cache.
6. Verify the message certificate.
When an entity wants to create a signed message 7. Validate the certificate chain.
for transmission, it has to contain the following 8. If needed, verify that the transmis-
information and services: sion location. Check if it is within the
GeographicRegion of the message signing
• A key for signing. certificate.
• The certificate associated with the signing 9. Verify the application field in the message
key. consists with that in the certificate.
• A random number generator. 10. Verify that none of the certificates in the
• A cryptographic implementation. certificate chain has been revoked.
• The current position and time; and the esti- 11. Verify the signature on each certificate with
mated error in that position and time. the public key from its issuing certificate.
12. Verify the signature on the message with
Then, the entity signs the message by taking the public key from the message signer’s
the following steps: certificate.
13. Cache any previously unseen certificates,
1. Fill the ToBeSigned Message structure their associated Application IDs and associ-
and encode it as the unsigned value octet ated Geographic Regions.
string.
2. Sign the unsigned message value. encrypted messages
3. Create and encode the signed message
value. The processes to encrypt messages are as fol-
lows.
An entity may receive a signed message. The
signed message contains: 1. Retrieve the certificate
2. Check that the recipient’s certificate has not
• The recently received message. been revoked
• The root certificate. 3. Select a symmetric encryption algorithm in
• The revoked certificate. the recipient’s certificate
4. Generate a random key for this algorithm
99
5. Retrieve and encode the message to be A RM in WAVE takes the role of a multiplexer,
encrypted directing the information exchange between RMA
6. Encrypt the message and RCP. WAVE specification aims at creating a
7. Create a RecipientInfo field completely interoperable communication envi-
8. Create and encode the EncryptedMessage. ronment through which vehicles and RSUs could
transfer data effectively.
If an entity wants to send a signed and encrypted
message, first it creates a signed message from the wave architecture
original message. Then, it creates an encrypted
message from the encoded SignedMessage. As shown in Figure 7, a RCP resides as a compo-
When an entity receipts a signed and encrypted nent in an OBU, while the RM can be a component
message, first it decrypts the message. Then, it of either the OBU or a RSU. RM and a resource
forwards the output message to the application manager application communicate via the security
for processing. tunnel in the between. RM, as a the multiplexer
in the WAVE framework, allows each RMA to
perform end-to-end communicate with RCPs on
waVe resource manager OBUs through itself, while the link between RM
and RCP is established through wireless (IEEE
This section describes the WAVE Resource 802.11p), which might be insecure.
manager, which is standardized in IEEE 1609, When a RM of a RSU sends respective mes-
specifying the wireless access method in a WAVE sage to a RCP on an OBU, the command would
environment. The proposed access method allows be executed by the RCP. Should the need arise; it
a RMA (remote manager application) to establish could response to the requesting RMA through the
connection with a RCP (resource command pro- RM. Therefore in Figure 7, RM provides services
cessor) on an OBU (on-board unit) through the for RMAs to access the memory and UIs on the
RM (resource manager) of a RSU (roadside unit). OBU, which are controlled by RCP. As shown in
Figure 7. Components addressed in IEEE 1609.1 (IEEE 1609.1TM, 2006)
100
Figure 8. Data transfer in WAVE
Figure 9. Four basic forms for services
Figure 7, if a RM resides in an OBU, this figure We will then talk about the data structure of
thus then depicts the situation of an inter-vehicle the message. When the RMA and the RM are not
communication session. in the same device, messages will be delivered in
packets, thus APDU; should they be in two stack
wave data Transfer protocols of a single identical device, either RSU or
OBU, we would then pass it as ASDU(application
As shown in Figure 8, the data transfer of RMA is service data unit). The messaging between inter-
done through command and response, which itself faces of different protocol layers is called SAP
includes a two phase action: the encapsulation of (service access point).
the information from RMA to RM containing in As shown in Figure 9, there are four types of
the APDU (application protocol data unit), and service indicated in this specification: Request,
the command/response as a whole from RM to Indication, Response, and Conform, each service
RCP. The steps are detailed below: can be in one of the following modes: Confirmed
Mode, Non-confirmed Mode, and Locally con-
1. RM received commands from a RMA; after firmed Mode (as indicated in Figure 10(a)~(c)).
confirmation and execution, the command
is send to the RCP of the OBU via IEEE application components
802.11p. Message from RM to RMA is
encapsulated in APDU, and those from RM In application layer, there are several components:
to RCP in UDP packets. OBU resources, RM commands and response, and
2. Command executed by RCP. Services provided by the RM.
3. RCP responds the execution result as well
as other information back to RMA through • OBU resources: OBU resource covers
RM. both Memory and User Interfaces:
101
Figure 10.Three modes for services
◦ OBU memory: OBU memory is di- 1), and transfer pages (type 2), which
vided into partitions, each further di- are detailed below:
vided into addressable blocks called ▪ Storage pages (type 0): General-
pages, which are accessed by RMA purpose pages, type 0 pages are
through RM command set. Since each controlled by applications, pro-
partition is indexed by a 16-bit integer, viding RMA data storage and
the largest partition is 64k bytes. In retrieval.
the same fashion, a largest page is also ▪ Memory-mapped pages (type
64k bytes. There must exist partition 0 1): OBU integrates many UI de-
in an OBU memory, while additional vices, which are controlled via
partitions are optional. Page 0 of parti- memory-mapped pages. These
tion 0 is reserved for special purpose, pages are treated as buffers for
and all access to this page would be devices. When a RMA want to
deemed invalid. OBU memory pages control a UI device, it simply
come in three types: storage pages writes the data to the page of
(type 0), memory-mapped pages (type that device. The command for
102
such action is Set User Interface indicates whether the receiver should
command. To get data from a reply, TRUE for no and FALSE
UI device, RMA executes Read for yes. Following 7 bits form the
Memory Page command. Command Transaction ID, which is
▪ Transfer pages (type 2): Transfer set by RMA when sending command
pages are the interfaces be- to RCP. When replying, this ID will
tween onboard equipments and be included without modification so
the network, and are related to that RMA could identify to which
RCP-controlled external inter- message it was responded.
faces. When RMA writes data ◦ Byte 3~4: This field is 2-byte long,
to a transfer page, the data entry used to indicate the length of the fol-
would be sent to the interface it lowing Command parameter field.
was related by. ◦ Byte 5~: Bytes after byte 5 is for
◦ User interface: User interface acts as the parameters which RMA sends to
bridges between machine and men. RCP.
Related GUIs are introduced below: ◦ Commands in resource management
▪ Visual display: Through which of WAVE are passed in the form of
an OBU could display message. a command sequence. When pass-
Acceptable colors in this speci- ing a single command, it would still
fication are red, green and blue. be transferred in a sequence with a
▪ Buzzer: Through an OBU could single entry. Sequences are processed
notify user of important events. in FIFO fashion once received by the
▪ Enunciator: Through which an client.
OBU can read messages. RM in WAVE would then process each
▪ Character readout: Through command sequence until a invalid com-
which an OBU can display text mand is met. While executing, if a FALSE
messages. No_Response_Indicator value is met, the
▪ Keypad: Through which a user RCP would need to send in a replying
can input message. response.
▪ Other future GUI
• RM commands and response: Command Shown in Figure 12 is the response format
format for RM is shown in Figure 11. Each of RM:
command is formed by 5 bytes, transferred
in sequence with the MSB of each byte • Byte 1: bit 7 is reserved, with the follow-
sent first. Fields in the command are ex- ing 7 bits forming the Command ID.
plained below: • Byte 2: bit 7 is unused, with the following 7
◦ Byte 1: Bit 7 is reserved; next 7 bits bits forming the Command Transaction ID.
form the Command ID, which is • Byte 3: this is the response status of RCP,
unique for each command. For infor example Command Success (0x01),
stance, the ID for Read Memory Page Command Failed (0x02), refer to IEEE
is 0x10. Refer to IEEE 1609.1 (2006) 1609.1 for more info.
for more information. • Byte 4~5: This field is 2-byte long, used
◦ Byte 2: Bit 7 is No Response to indicate the length of the following
Indicator, which, for its namesake, Response Data. Note that this field and the
103
Figure 11. Command format. (IEEE 1609.1TM, 2006)
following one are used only when making • Reserve Partition command (ID: 0x43)
reply for Read Memory Page command. • Release Partition command (ID: 0x44)
• Byte 6~: Response Data
• Services provided by the RM: RM takes
Following IDs are further specified in IEEE the role of the application layer in WAVE
1609.1 (2006): by offering related services to RMA, which
are either protocol management services
• Read Memory Page command (ID: 0x10) or protocol data transfer service, which
• Write Memory Page command (ID: 0x11) we are about to detail here. Each service
• Insert Message command (ID: 0x12) has a service name in the form of RMA-
• Sleep Transaction command (ID: 0x30) <SERVICE>-<ACTION>, for example
• Reserve Memory Page command (ID: 0x40) RMA-ACTIVATE-Request.
• Release Memory Page command (ID: 0x41)
Figure 12. Response format. (IEEE 1609.1TM, 2006)
104
◦ Protocol management servic- send to RM a RMA-

es: Protocol management services TERMINATESESSION-Indicate.
cover RMA-ACTIVATE-Request, ▪ Terminate session con-
R M A - A C T I VAT E - R e s p o n s e , firmation service: Named
RMA-NOTIFY-Indication, RMA- RMA-TERMINATESESSION-
NOTIFY-Confirmation, RMA- Confirmation. RM would
TERMINATESESSION-Indicate, notify RMA through RMA-
RMA- TERMINATESESSION- TERMINATESESSION-
Confirmation, RMA-DEACTIVE- Indicate that RMA-
Request, and RMA-DEACTIVATE- TERMINATESESSION-Indicate
Response, which has been received.
▪ Activate request service: ▪ Deactivate request service:
This service is named RMA- Named RMA-DEACTIVE-
ACTIVATE-Request. Once the Request. When RMA wishes to
RMA is ready to communicate stop the operation, it would send
with RCP, it notifies the RMA this service to notify RM.
through the RMA-ACTIVATE- ▪ Deactivate response service:
Request service provided by Named RMA-DEACTIVATE-
RM. Response. When RM received
▪ Activate response service: R MA -DE A C T IV E -R eq u e s t ,
This service is named RMA- it would respond with RMA-
ACTIVATE-Response. When the DEACTIVATE-Response.
RM has completed its activation, ◦ Protocol data transfer services:
it gives a response to RMA. Protocol data transfer services covered
▪ Notify indication service: This RMA-EXCHANGE-Request, RMA-
service is named RMA-NOTIFY- EXCHANGE-Response, and RMA-
Indication. Once RMA is acti- EXCHANGE-Confirmation:
vated, it would wait for its part- ▪ Exchange request service:
ner. Through this service, RMA Named RMA-EXCHANGE-
can be informed once the RCP Request. RMA uses to send com-
of a OBU has entered the command sequences to RCP.
munication zone of the service ▪ Exchange response service:
provider. Named RMA-EXCHANGE-
▪ Notify confirmation service: Response. Once command se-
This service is named RMA- quence is sent to RCP and the re-
NOTIFY-Confirmation. Once the sponse sequence is received, RM
RMA received RMA-NOTIFY- would reply to RMA with RMA-
Indication, it would respond with EXCHANGE-Response.
a RMA-NOTIFY-Confirmation. ▪ Exchange confirmation ser-
▪ Terminate session indica- vice: Named RMA-EXCHANGE-
tion service: Named RMA- Confirmation. When RMA gets
TERMINATESESSION- the response of RM, this message
Indication. When a transaction would be sent back to RM.
is completed, RMA would
105
reFerences IEEE1609.3TM (2006). IEEE Trial-Use Standard

for Wireless Access in Vehicular Environments
IEEE1609.1TM (2006). IEEE Trial-Use Standard (WAVE) – Networking Services.
(WAVE) – Resource Manager. IEEE1609.4TM (2006). IEEE Trial-Use Standard
IEEE1609.2TM (2006). IEEE Trial-Use Standard (WAVE) – Multi-channel Operation.
(WAVE) – Security Services for Applications and
Management Message.
106
Section 3
108
Chapter 7
Introduction to Global Satellite
Positioning System (GPS)
Jenq-Muh Hsu
National Chiayi University, Chiayi, Taiwan, R.O.C.
absTracT
Understanding the right positions and directions of people and objects is a significant issue from the
ancient eras to the present. In the past, people often launched a war in order to satisfy the craving for
the dominating powers and spread their realms. In the recent, Global Satellite Positioning System (GPS)
has become the one of most popular positioning technologies. GPS can provide users precise position-
ing information, no matter wherever that may present their own positions. The early GPS positioning
technology has been widely used in military, marine use, until recently gradually applied into our daily
life, e.g., automotive navigation, geodesy surveying, etc. In this chapter, we will briefly introduce some
GPS issues including the origins of GPS, GPS system architecture, and related GPS applications.
inTroducTion South Pointing Chariot (South Pointing Chariot,

2009) invented in the ancient Chinese civilization
Understanding the right positions and directions of is a complex-gearing mechanism without using
people and objects is a significant issue from the the magnet to point the same direction while the
ancient eras to the present. In the past, people often chariot is around in any movement. South Point-
launched a war in order to satisfy the craving for the ing Chariot only moves on the ground, it can not
dominating powers and spread their realms. When be used to point the right direction whiling sealing
the army is marching during the fighting, it does due to the no effect of gearing operation on the sea.
not allow the army to get lost in the woods or the The other usual positioning and navigating tool is
dense fog. It needs a method or a tool to offer the the compass.
army to navigate and move to the right place. The compass has the magnet which can interact
with the magnetic field of the earth to always point
DOI: 10.4018/978-1-60566-840-6.ch007 the same position in north. Although the actual
Introduction to Global Satellite Positioning System (GPS)
positioning is a magnetic North Pole, not a true the GPS receivers to receive enough GPS signals
North Pole, it is enough to use in positioning without bidirectional communications among
and navigation. Therefore, the compass has been them. GPS receives will be able to calculate out
widely used in many positioning and navigating the user positions they locate now.
applications. The compass still has a little ac- GPS system generally consists of a constella-
curate positioning problem affected by magnetic tion of 24 operational satellites (Leick, 2004). In
interference in natural environment, but it has no order to ensure continuous worldwide converge
serious influence on using the compass. for GPS positioning service, six orbital planes of
Nowadays, the most famous positioning and satellites are organized and each four satellites
navigation service is Global Satellite Position- are placed in an orbital plan. There are four to
ing System (GPS) (Global Positioning System, ten GPS satellites will be visible anywhere in
2009; EI-Rabbany, 2002; Kaplan et al., 2006). the world under this constellation geometry. The
It uses the satellites orbiting around the earth in sketch map of GPS constellation is shown in Fig-
space to broadcast positioning signals. A GPS ure 1. A GPS satellite routes around the earth in a
receiver receives the signals and calculates out its nearby circular orbit, an elliptical shape, with an
current position. That is, GPS provide a world- inclination of 55 degrees to the equatorial plane.
wide positioning and navigation service for any The maximal radius of GPS orbit is about 26,560
kind of communication and transportation, such kilometers measured from the earth center. The
as aircrafts, vessels, vehicles, pedestrians, etc. orbit period of GPS satellite is approximately 12
GPS is developed by the U. S. Department of sidereal hours, which is about 11 hours and 58
Defense (DoD) in early 1970s. At the beginning minutes. Thus, GPS satellite will run around the
of constructing GPS, it is only used for military earth twice per day. In order to ensure the avail-
purpose, but it has freely opened for civilian use ability of GPS positioning service, the number
on July 2000s. Therefore, GPS is now a dual-use of satellites in the GPS constellation has always
system that can be accessed by both military and been more than 24 operational satellites.
civilian users in various positioning precision
for some national security factors considered by
U.S. government. Figure 1. The GPS constellation
As mention above, vehicles having built-in
GPS navigation system will be a main trend to
provide the driving assistance and navigation ser-
vice in the future. In next section, we will describe
the principle of GPS system in detail.
principle oF global
posiTioning sysTem
The GPS is a satellite-based positioning and

navigation system. GPS provides continuous po-
sitioning and timing information anywhere in the
world under any weather conditions. GPS is also a
passively one-way ranging system and it can serve
unlimited number of users. That is, users only take
109
The GPS system is comprised of three segments ground antennas shown in Figure 3. The
shown in Figure 2: space (satellite constellation), main task of control segment is responsible
control (ground-control/monitoring network), and for tracking, monitoring, commanding, and
user (receiving equipment). The detailed functions controlling the GPS satellite constellation in
of each segment are listed as follow: order to determine and predict satellite loca-
tions, system integrity, behavior of satellite
1. Space segment: It consists of 24-statellite atomic clocks, and other considerations. The
constellation. It also has some back-up satel- observation of GPS satellites collected at the
lites to ensure the full operational capability monitor stations are transmitted to the MCS
(FOC) of GPS positioning service. Each for processing. MCS sends the processing
satellite transmits a unique PRN (Pseudo outcome, the fresh navigation data (control
Random Noise) ranging signal to measure information), to the monitor control stations
the distance from GPS receiver to GPS satel- with ground antenna uploading them to GPS
lite. The satellite signals are controlled by satellites through S-band link. Therefore,
atomic clocks with high precision to transmit the control segment is an important role to
through two L-bands, L1 (1,575.42 Hz) and maintain the operational behavior of GPS
L2 (1,227.6 Hz). GPS satellite has an S-band positioning service.
antenna to communicate with control seg- 3. User segment: It typically refers to as a GPS
ment in order to maintain the operational receiver shown in Figure 4. When a user uses
control of GPS service. a GPS receiver to receive enough the L-band
2. Control segment: It consists of a master signals from GPS satellites under a spacious
control station (MCS) and a worldwide area, the GPS receiver can determine user’s
network of five monitor stations and four position and related information, including
Figure 2. GPS segments
110
Figure 3. Geographic distribution of control segment facilities
Figure 4. Various types of GPS receivers
longitude, latitude, altitude, time, velocity, distances from a point holding a GPS receiver
etc, anywhere in the world. on the surface of the earth to three GPS satellites
are known along with satellite positions, the posi-
tion of the point can be determined by triangular
gps posiTioning model positioning theory. The concept of TOA ranging
model is to measure the propagation time of signal
The basic idea of GPS positioning model is very broadcast from signal emitter at known locations.
simple. It uses the TOA (Time of Arrival) rang- Assume that the signal emitter regularly broadcast
ing model to measure user position. That is, the the beacon signal and both of signal emitter and
111
receiver have synchronous clocks. The receiver foghorn whistle is repeatedly sounded per minute,
can measure the propagation time of beacon and the position of foghorn is known. The marine
signal from emitter and then propagation time is measures the elapsed time from minute mark until
multiplied by the speed of the signal to obtain the the foghorn whistle is heard. Thus, propagation
distance from the emitter to the receiver. time (t) of foghorn whistle multiplies speed (v) of
GPS system provides three-dimensional po- the sound is the distance (R, R = v × t) from the
sition information including longitude, latitude, foghorn to the marine. Let the foghorn be denoted
and altitude. It needs to use the three-dimensional F1 and the distance from the foghorn F1 to the
position determination to calculate the position of marine be denoted R1 respectively. Thus, marine
the measured position. In order to simply examine only knows the vessel is somewhere on a circle
the GPS positioning model, we first introduce the with radius R1 centered about the foghorn, which
two-dimensional position model how it achieve is shown in Figure 5-(1). If there is the second
the point positioning. Consider a mariner at sea foghorn F2 and the marine can simultaneously
determine his vessel’s position from a foghorn. measures the distance R2, the vessel position will
Assume that the vessel and the foghorn are be located at one of the intersections of the range
equipped with accurate synchronized clocks, the circles with radius R1 and R2 centered about the
Figure 5. Two-Dimensional Position Determination: (1) All possible positions ranging from a single
source, (2) Two ambiguous positions from measurements to two sources, and (3) Position ambiguity
removal by three sources.
112
Figure 6. GPS Three-Dimensional Positioning

foghorns F1 and F2 respectively, which is shown
Model: (1) User located on surface of sphere, (2)
in Figure 5-(2). In order to resolve the position-
User located on perimeter of plane of intersection,
ing ambiguity of the vessel, third foghorn F3 is
and (3) User located at one of two points.
applied to measure the distance R3 for removing
position ambiguity as shown in Figure 5-(3). In
fact, the TOA measurement would not be perfect
due to the clock drift, sound interference, atmo-
spheric effects, etc. These errors would affect the
measurement and resulting in accurate distance
computations.
GPS also employs TOA ranging and three-
dimensional positioning model to calculate user
position. The measurement of GPS positioning
model is similar to two-dimensional positioning
model. Assume that each satellite transmits a
unique ranging signal in space. A satellite has a
clocks synchronized with others clocks built on
other satellites. The clock synchronization and
orbit plane of these satellites are monitored and
controlled by the control segment (MCS). A GPS
receiver also has a clock synchronized to GPS
system time. Ranging signals transmitted from
GPS satellites contain time information. It enables
the receiver to measure the propagation time (t1) order to measure the more accurate user position,
of the ranging signal left the satellite (S1) based the fourth satellite will be applied to calculate and
on satellite clock time. Thus, the satellite-to-user fix the user’s altitude.
distance (R1, R1 = Vlightspeed × t1) can be calculated. In addition GPS can determine the user veloc-
The user would be located somewhere on the ity. The most popular used method is based on
surface of a sphere centered about S1, as shown in estimating the Doppler frequency of the received
Figure 6-(1), in this ranging measurement. If the GPS signals. GPS also can measure the moving
second satellite (S2) is considered into account, the orientation (direction) of a rigid object, e.g., a
user would be located somewhere on the surface vessel, an aircraft, etc.
of both spheres ranging from S1 and S2, as shown
in Figure 6-(2). That is, the user position is located
at a plane of intersection. The third satellite (S3) gps reFerence
is adopted to measure the user position. The user coordinaTe sysTem
would be located at the interaction of the perimeter
of the circle and the surface of the sphere centered Although GPS can measure the user position
about S3, as shown in Figure 6-(3). We can find anywhere, under any weather conditions. Due to
that the user position would be at two possible the irregular topographic surface of the earth, the
points, and only one of the points is the correct determination of the user position is very difficult.
user position. For the user position, it is generally To over this problem, it needs a mathematical
locates on the surface of the earth. Therefore, the formulation to model the smooth surface of the
lower one of two points will be the true position. In
113
earth. Therefore, it needs a reference coordinate (λ), and the altitude (h) above the reference surface.
system to represent the states of GPS satellites That is, geodetic coordinate (ϕ, λ, h) of Point P
and the receiver. A coordinate system is a set of can be easily transformed to or from Cartesian
rules to specify the locations of points. This usu- coordinate (x, y, z).
ally specifies the origin (the central) and a set of The world geodetic system (WGS) (World
reference lines (the axes) with known orienta- Geodetic System, 2009) is a standard coordinate
tion. Figure 7 shows a 3-D Cartesian coordinate system used in cartography, geodesy, and naviga-
system which uses three reference axes, x, y, and tion. It consists of a standard coordinate frame for
z, intersecting at the origin (C) of the coordinate the earth, a standard spherical reference surface
system. Coordinate systems can be classified for raw altitude data, and a surface (nominal sea
according to the reference surface, the origin, level) with the same gravitation. That is, The WGS
and the orientation of the axes. For example, in can provide an ellipsoidal model to represent the
a 3-D geodetic coordinate system (Torge, 1991), real earth shape. Therefore, the WGS can provide
the reference surface is an ellipsoid. The orienta- the means for relating positions on various local
tion of the axes and the origin are specified by geodetic systems to an Earth-centered, Earth-fixed
two planes: the meridian plane through the polar (ECEF) coordinate system. A series of WGS
and equatorial plane. It is important to provide an revolution, WGS 60 (developed in 1960), WGS
easy-understanding positioning information for 66, WGS 72, and WGS 84, have been developed
the GPS users. Thus, positioning information in a to provide more accurate coordinate system. Cur-
3-D geodetic coordinate system can be described rently, the GPS system uses the WGS 84 to be the
in the geodetic latitude (ϕ), the geodetic longitude standard of coordinate reference system.
Figure 7. A sample of 3-D coordinate system
114
gps daTa proTocol FormaT all GPS receivers should support this speed. But
some modern GPS receivers can support higher
Different GPS manufacturers may have their interface speeds, e.g., 9600 or 38400 bits per
own data formats to store the GPS measuring second, to communicate.
information. It may lead to increase the difficulty Data stream in the NMEA 0183 standard is an
of system integration. Thus, it is necessary to ASCII format. It includes the complete position,
define the standards of GPS data formats to ac- velocity, and time (PVT) information. The data
cess the uniform GPS information from various is sent in the form of sentences. Each sentence
GPS receivers. A number of standard formats of begins with a ‘$’ character and ends with a car-
GPS information representation for various needs riage return/line feed <CR><LF> sequence. The
have been developed, which includes RINET, maximal length of sentence is less than 80 char-
NGS-SP3, RTCM SC-104, NMEA 0183, etc. acters of visible text. The data fields in sentence
NMEA 0183 is the most popular GPS data format formed in a single line are separated by commas.
in these standards. The first 5 letter prefix following the beginning
The National Marine Electronics Association character ‘$’ identify the sentence type (first 2 let-
(NEMA) has developed a specification defining ters) and sentence content (successive 3 letters) in
the standard, called NMEA 1083, (NMEA Data, a sentence. The prefix of sentence type is GP for
2009) communicating interface among various GPS receivers. The last field in any sentence is
marine electronic equipments and application a checksum field following a delimiter character
software. The standard allows marine electron- ‘*’. NMEA standards have defined various GPS
ics to send information to computer or to other sentences. Table 1 shows some common GPS sen-
marine equipment. GPS receiver communica- tences used in a GPS receiver and an example for
tion is also defined within the NMEA standard. decoding GPGGA sentence is illustrated in Table
GPS receivers generally use the RS232 protocol 2. Besides, we also can visit the web site (http://
(EIA-422) to communicate with GPS software gpsinformation.org/dale/nmea.htm) to view the
through computer serial ports. The interface detailed structures and formats of other common
speed of NMEA standard is 4800 baud rate with GPS sentences.
8 bits of data, no parity, and one stop bit. Thus,
Figure 8. An example of valid GPS sentences generated from a GPS receiver
115
Table 1. Some common GPS sentences used in

GPS information can be visualized displayed on
GPS receiver
the GPS viewer software or equipment.
Name Notes Figure 9 depicts an example of GPS viewer
GPGGA Global positioning system fixed data
showing GPS information including the longitude,
GPGSA GPS DOP and active satellites
latitude, height, speed, direction, the number of sat-
GPGSV Satellites in view shows data about the status of
ellites in view of the sky, their locations of viewed
satellites satellites, and the strength of GPS signals.
GPRMC Recommended minimum command
GPVTG Velocity information
GPDLL Geographical Latitude and Longitude gsp applicaTion For
Vehicle naVigaTion
Figure 8 shows a series of GPS sentences Since the price of GPS receiver has become
generated from a GPS receiver, which sentences cheaper, the usage of GPS has also become more
contain five main content types: GPGGA, GPGSA, ubiquitous. Thus, many innovative GPS services
GPGSV, GPRMC, and GPVTG. After decoding a including navigation and positioning, tracking,
series of GPS sentences, which include GPGGA, and surveying, etc., have been widely applied into
GPGSA, GPGSV, GPRMC, and GPVTG, these various application domains. In this section, a GPS
Table 2. The decode of GPGGA sentence
GPGGA Sentence $GPGGA,040905.000,2333.2209,N,12025.6358,E,1,07,1.2,37.6,M,17.1,M,,0000*61<CR><LF>

Data Items in Sentence $GPGGA,[1],[2],[3],[4],[5],[6],[7],[8],[9],[10],[11],[12],[13],[14],*[15]<CR><LF>
Number Notes Example

[1] Time of position in UTC system [hhmmss.ss] 040905.000
[2] Latitude [lll.ll] 2333.2209
[3] North or South [N/S] N
[4] Longitude [yyyy.yy] 12025.6358
[5] East or West [E/W] E
[6] GPS quality indicator [0=invalid; 1=GPS fix; 2=Diff. GPS fix, …] 1
[7] Number of satellites in use [xx] 07
[8] Horizontal dilution of position (HDOP) [x.x] 1.2
[9] Antenna altitude above/below mean sea level (geoid) [x.x] 37.6
[10] Meters (Antenna height unit) [M] M
[11] Geoidal separation (Diff. between WGS-84 earth ellipsoid and mean sea level. -=geoid is below WGS-84 ellipsoid) [x.x] 17.1
[12] Meters (Units of geoidal separation) [M] M
[13] Age in seconds since last update from diff. reference station [x.x] (NULL)
[14] Diff. reference station ID# [0000-1023] 0000
* Checksum delimiter character [*] *
[15] Checksum (hh) 61
<CR><LF> Sentence terminator [<CR><LF>] <CR><LF>
116
Fgiure 9. An example of GPS viewer virtualizedly

vehicle GPS navigation system (automotive GPS
showing GPS information
navigation system) can assist the driver to follow
a route plan to reach the driving destination s/he
sets. Figure 10 shows an example of using GPS
navigation to be an aid for driving during a trip.
Although GPS navigation system can acquire
the position information from GPS satellite sig-
nals, it needs a digital map stored in the system
or retrieved from the Internet to locate the user
or the car on the road in a road network. Thus, a
newest full set of digital map is a very significant
in order to provide an accurate GPS navigation for
matching to the real road network environment,
e.g., a new road has been opened. A friendly user-
interface for user is also an important issue in a
GPS navigation system. A modern GPS navigation
system may consist of visual-display interface
(e.g., virtual road network in 3D view), touch
screen (e.g., target input), voice-based interface
application concentrating on vehicle navigation (e.g., speech recognition and synthesis), etc., to
will be discussed. satisfy the various requirements of the GPS driving
GPS provides the location, velocity, orienta- guidance. Some GPS navigation systems may have
tion, and time information that enables many a module, to receive the real-time traffic informa-
practical applications to be used in our daily lives. tion to dynamically update the navigating guidance
The first popular application is GPS navigation service fitting the real road network condition via
service (Drane et al., 1998). A car installed an in- traffic message channel (TMC) (Traffic Message
Figure 10. Use the GPS navigation system to aid the drive
117
Channel, 2009) through FM radio broadcasting, Differential, G. P. S. (n.d.). Retrieved September

WiFi/GPSR/3G data transmission, inter-vehicle 11, 2009, from http://en.wikipedia.org/wiki/Dif-
communication, etc. ferential_GPS.
In addition to GPS navigation service, the GPS
Drane, C., & Rizos, C. (1998). Positioning system
can also be used in vehicle tracking, GPS security,
in intelligent transport systems. Boston: Artech
fleet management, intelligent public transportation
House.
service, emergency rescue, biological movement,
geodetic surveying, graphic information system, EI-Rabbany. A. (2002). Introduction to GPS –
location-based service, etc. We also find that The Global Positioning System. Boston: Artech
a cellular phone with a built-in GPS receiving House.
module is a trend. It means that cellular phone
Global Positioning System. (n.d.). Retrieved
will act a portable and personal GPS location and
September 11, 2009, from http://en.wikipedia.
navigation platform to provide the walker navi-
org/wiki/Global_Positioning_System.
gation with highly accurate guiding information,
the location-aware services of the four essential Kaplan, E. D., & Hegarty, C. J. (2006). Under-
requirements of the people: food, clothing, hous- standing GPS – Principle and Applications.
ing and transportation, in the future. Boston: Artech House.
Leick, A. (2004). GPS Satellite Surveying. New
York: John Wiley & Sons.
conclusion
NMEA Data. (n.d.). Retrieved September 11, 2009,
In this chapter we have briefly introduced an from http://gpsinformation.org/dale/nmea.htm.
essential principle of the GPS system, including
South Pointing Chariot. (n.d.). Retrieved Septem-
GPS constellation, GPS segments, GPS position-
ber 11, 2009, from http://en.wikipedia.org/wiki/
ing model, standard GPS data format, and GPS
South_Pointing_Chariot.
application for vehicle navigation. There are still
many issues, such as GPS satellite signal char- Torge, W. (1991). Geodesy. Berlin: Walter de
acteristics and acquisition, GPS signal inference Gruyter.
and multipath, error measurement, differential
GPS (D-GPS) (Differential GPS, 2009), assisted Traffic Message Channel. (n.d.). Retrieved Sep-
GPS(A-GPS) (Assisted GPS, 2009), other GPS tember 11, 2009, from http://en.wikipedia.org/
systems, geodetic/geocentric coordinate system, wiki/Traffic_Message_Channel.
map projection, etc., to be discussed. World Geodetic System. (n.d.). Retrieved Septem-
ber 11, 2009, from http://en.wikipedia.org/wiki/
World_Geodetic_System
reFerences
Zhao, Y. (1997). Vehicle Location and Navigation
Assisted, G. P. S. (n.d.). Retrieved September Systems. Boston: Artech House.
11, 2009, from http://en.wikipedia.org/wiki/As-
sisted_GPS.
118
119
Chapter 8
Vehicle Location and
Navigation Systems
Ben-Jye Chang
National Yunlin University of Science and Technology, Yunlin, Taiwan, R.O.C.
absTracT
The most driving purpose is to traverse to the destination safely, efficiently, and comfortably. Two types
of approaches could achieve the goals, including the static and dynamic approaches. In the static aspect,
vehicles use the static road and traffic information to navigate. Conversely, in the dynamic aspect, ve-
hicles adopt the dynamic information instead. However, both of the two approaches first require getting
the vehicle’s location and then map the position on an e-map. Thus, this chapter first introduces some
important vehicle location determination algorithms: the dead reckoning and global position system
algorithms, in which the precision of location technologies are compared. Then, the map-matching
algorithm is described in detail. Finally, various vehicle navigation approaches are detailed, in which
the important topics include: the navigation architecture, the navigation routing algorithm, and naviga-
tion applications.
inTroducTion only waste our time, but social cost. Therefore, trans-
portation simplification, driving time reduction, and
The invention of the vehicles, though it could shorten energy conservation are the top priority concern of
the distance from place to place, has brought a lot the world. Thus, this chapter, Vehicle Location and
of serious issues: traffic jam, car accident, energy Navigation Systems (Farrell et al., 2008; Zhao et al.,
crisis, environmental pollution, etc. Moreover, as 1997), are what we researched. While the vehicles
the population rises and living in standard increases are moving, realizing your location, destination and
constantly, the vehicles are everywhere and the roads the shortest path are important. However, most of
are even more complicated. Finally, it would not all, you know where you are, or other conditions
are useless. The early days of vehicle position are
DOI: 10.4018/978-1-60566-840-6.ch008 a driver reads maps and road signs in the meantime
Vehicle Location and Navigation Systems
Figure 2. Distance sensor device

and finally reaches its destination. It is inefficient
and dangerous still. Moreover, wasting resources
and other problems could occur by vehicles posi-
tion. Generally, vehicle position can be divided
into two methods: early days’ dead position and
current used GPS technologies.
For locating a vehicle position, the vehicle with
dead reckoning method must set its coordinates
before moving. Position and distance sensors,
based on the coordinates, could figure out tracking
movement and turning degree. Lastly, the current
position could be located.
For example, as shown in Figure 1, the starting
point (x 0 , y 0 ) , the based point also, multiplied by
However, the results could be a huge error due to
each distance di and aspect qi is the terminal point different wheel sizes and terrains (e.g., idle running
(x n , yn ) . The formula becomes as follows: of wheels on the snow and sand). Distance sensors
are subdivided into mechanical and electronic
n
x n = x 0 + å diT cosqi ones. Figure 2 is a typical example.
i =0 The vehicle position sensors enable position
n
measurement. It can either be an absolute position
yn = y 0 + å diT sinqi (1)
i =0
sensor or a relative one. The first one applies the
earth’s magnetic field to get absolute position.
Distance and position sensors are applied by
Figure 3 shows that the induction coils can be
dead reckoning method. The following contents
affected by the earth’s magnetic field and the
in this book would discuss both two common
vehicle’s position can be measured can by the
sensors.
angle q , as demonstrated in equation 2. However,
Distance sensor normally applies speedometer
using earth magnetic field to induct potion has a
to plot the moving distance. The application makes
disadvantage. That are the results are easily af-
use of the sensor device, which can calculate wheel
fected by some circumstances: viaducts, bridges,
turns, on the wheel axle. Circumference of the
substation, and high-voltage tower. Conversely,
wheel multiplied by the wheel turns is the distance.
the relative position is determined in terms of the
original point and angle.
Figure 1. Dead reckoning
Vx
q = tan-1 (2)
Vy
Dead reckoning was mainly used in the past
with three disadvantages: 1) Based point should
be located well, 2) The sensor has huge error,
and 3) The more distance you move, the more
error could be.
120
Figure 3. Earth magnetic sensor
global positioning system Conversely, some disadvantages of GPS

still need to be addressed. For example, if a cell
Since the first generation GPS chip was presented phone or radio’s wave band is closed to GPS’, it
to the public, it overcomes some problems: un- would cause interference. In addition, tunnels and
clear signal of ex-chip, long positioning time, weather would affect the accuracy of GPS also.
location drifting, etc. Due to the progress of chip Thus, adjusting GPS error by a fixed base
manufacturing procedure and size reduction, even station and then conveying it to the vehicles’
PDA, digital camera, cell phone and so on can be GPS for correcting can add accuracy. It is called
integrated with GPS. The GPS device equipped Differential GPS (namely D-GPS). As Figure 4
on the vehicles is simple and more accuracy than demonstrated, the FM radio can also receive the
dead reckoning method. It is getting to be used as positioning signal from the base station.
basic vehicle equipment.
Figure 4. Differential GPS (D-GPS)
121
map maTching al., 2005). In a digital map, a road is viewed as a

link and an intersection is viewed as a node, as
The vehicle positioning system is different from shown in Figure 6.
the airplane and ship positioning systems. The Furthermore, when a vehicle is moving, the
reason is that the vehicle movement tracks must on coordinates of the vehicle can be compared with
roads. Recently, maps have already been digitized; the digital map, as demonstrated in Figure 7. For
thus, the error between digital maps and move- example, the black bold line, namely a road, and
ment tracks can be corrected by computers. As a the green one is the vehicle movement track, in
result, a more accurate position can be resulted, which we could get two vertical lines: L1 and
as shown in Figure 5. L2. If L1 is shorter than L2, it means the vehicle
movement track is on road A, and vice versa. This
map matching method is the basis theorem of the map matching.
Although the vehicle positioning can use map
The theorem of map matching is to match a vehicle matching to correct the movement tracks and to
moving track on a digital map (Brakatsoulas et get accurate position, map matching still has some
Figure 5. Map matching
Figure 6. Digital network map
122
Figure 7. Map matching
drawbacks. For example, if a road is just not on GPS, the vehicle position can be opened another
the digital map or one drives to parking lots or different applications, e.g., the vehicle tracking,
open ground, how to calculate position with map the transportation online information, etc.
matching? Therefore, an error upper and lower
range should be set. Once an error is in the range, introduction of Vehicle Tracking
it initiates the correction of the movement tracks;
otherwise, the error can be ignored. The vehicle tracking issue is a vehicle equipped
with GPS and has the capability to report its co-
ordinates to a computer periodically. In Figure
Vehicle Tracking 8, a user can trace the target vehicle’s position
through the Internet. Similarly, more and more
The GPS device is easy to be carried and easy new applications of vehicle tracking are being
to be integrated with other systems. Moreover, created, for example, a remote alarm system,
due to the common used and free charge of using transportation tracking, etc.
Figure 8. Vehicles tracking
123
model of the Vehicle effectively to avoid embezzlement and be

Tracking system managed easily.
• Car alarm system: The best car alarm sys-
Basically, the model of the vehicle tracking system equipped also by car rental agencies
tem can be divided into two parts: the position now is the vehicles tracking system. Once
dispatch end and the reception end. In the posi- the vehicle was stolen, we can find out no
tion dispatch end, the GPS receiver continuously matter wherever it is.
receives the satellite signal and then locates the
vehicle’s position. Meanwhile, the coordinates can The distribution and management system of
be dispatched to the Internet through the wireless police cars and ambulances through surveillance
network base stations. The position reception end center management. If something urgently hap-
can receive and match the coordinates from the pened, they can send the nearest police cars or
Internet and pass current position to the digital ambulances to support. Furthermore, the patrolling
map. Finally, the date can be shown on the screen, position and the public security can be improved
as demonstrated in Figure 9. efficiently.
The applications of the Transportation:

Vehicles Tracking
• Cargo services: Through the vehicles
The applications of the vehicle tracking are very tracking system, we can know cargo des-
extensive. It is even more than what we can think tination, map out a route to avoid wasting
about of. Till now new applications are still brought time. A consumer can realize the cargo’s
up. However, normally they can be divided into status through the Internet.
several parts.
The delivery of dangerous goods and polluted
Security: material, e.g., through the vehicle tracking system,
once chemical or toxic material, stone delivery
• Cash-carrying vehicle security surveil- and the like are leaking out, we can deal with it
lance: So far the vehicles tracking sys- in the very beginning and avoid other vehicles
tem, a basic equipment for a cash-carry- entering the limited area.
ing vehicle, can track a vehicle’s position
Figure 9. Fame of the vehicles tracking system
124
Table 1. The precisions of different location

Public transit:
technologies
• Vehicles distribution: Before you head Technology Medium Precision
somewhere, through the Internet, one can GPS Radio 25 m
check the nearest taxi or bus current posi- DGPS Radio 3m
tion. Moreover, through the vehicle track- WAAS Radio 3m
ing system, speeding and route time can be
MPS Radio 150 m
estimated.
• The management of a fleet of vehicles: It is
not only for making up for cell phone using
coordinates computed by a location technology
defects, but also avoiding lagging behind
may exist significant erroneous. By integrating
and getting lost.
with several location technologies is a good ap-
proach for reducing errors.
auTonomous locaTion
and naVigaTion
autonomous navigation
In the past, the cost of navigation system was very
Navigation means that a position can be located
high, so that it was used on special applications.
on a map and a routing algorithm referring to
In increase of time, the cost is down and drivers
the location of the vehicle calculates the path
use the navigation system to locate the position
for a driver. Autonomous navigation means that
and to navigate without any help. In this section,
navigation is done by self. Figure 10 proposes an
we briefly depict the issues of the autonomous
example of the autonomous navigation architec-
location and the navigation systems.
ture (Zhao, 1997). The communication devices are
used to receive the radio signals and to calculate
autonomous location
the position. After determining the position, the
user can input the destination for the system to
The autonomous location is that a vehicle can
determine the routing path.
locate the position by itself. In (Zhao, 1997; Zhao
et al., 1997), authors classified the location tech-
nologies to three types: the stand-along location
technology (Chang et al., 2008), the terrestrial Figure 10. The architecture of a navigation
radio technology and the satellite technology. In system
the stand-alone location technology, the Dead
Reckoning is an example to use the self ability
to locate the position. The terrestrial radio tech-
nology and the satellite technology are similar,
because the terrestrial radio technology adopts the
signals of base stations to locate the position and
the satellite technology uses the satellite signal
to locate position.
The precisions of different technologies are
indicated in Table 1 (Schiller et al., 2004). Based
on the results, only using one location technology
is not enough for some applications, because the
125
A navigation system not only provides the rout- centralized architecture

ing function, but also offers additional functions
such as weather bulletins, restaurant position, etc. Zhao (1997) shows the Centralized architecture,
Based on the factors, designing a navigation sys- as shown in Figure 11. Communication devices
tem must consider what typical kinds of functions are adopted to communicate with the wireless
should be included in the navigation system. technology such as WiMAX and GSM in order to
perform the communication between the client and
the server. A client requests a service from server
cenTraliZed locaTion and the server provides a service to the client.
and naVigaTion
location architecture
After describing the navigation and location
technologies in the previous section, this section In Figure 12, Zhao (1997) introduced a location
details the location and navigation systems based architecture of a centralized system. This archi-
on a centralized framework. In the centralized tecture consists of three parts including commu-
location and navigation systems, all functions nication devices, a client and a server. The client
are done in a centralized server for reducing the uses the GPS device or other methods to get the
computing loading of clients. In this section, we coordinates, and then sends it to the server. When
first introduce the architecture of centralized receiving the coordinates of the client, the server
systems. Location architecture and navigation computes the position and locates it on a digital
architecture are then depicted. map. The final result is shown on a graph-based
interface, as demonstrated in Figure 13.
Figure 11. System relation between communication devices, client and server
Figure 12. Location architecture of a centralized system
126
navigation architecture these technologies that automatic vehicle needs.

Firstly, we introduce automatic vehicle driving and
In Figure 14, Zhao (1997) introduces a navigation navigation. The needs of technologies of automatic
architecture of a centralized system. In the client, vehicle are then discussed in detail.
after keying the destination, the client sends a
navigation request to the server. When the server introduction
receiving a request, the server determines a path
for the client and then sends the determined rout- In the past, an automatic vehicle was a dream but
ing path to the client. According to the received now technologies are advanced and can be accom-
results, the client drives to the destination via the plished. The traffic accidents are often happened
central-computed routing path. when drivers are scatterbrained or other factors
which are not avoided. By using the automatic
vehicle, it could reduce the probability of accident.
auTomaTic Vehicle driVing An automatic vehicle also saves us a lot of time
and naVigaTion which we could use to do other things.
For realizing an automatic vehicle that can
Automatic vehicle driving technology let user automatically drive to a destination, some key
drives the vehicle without any operations. How to technologies such as image process, navigation
achieve this goal? In this section, we thus discuss system and automatic control system are required.
Figure 13. (a) A scenario of simulation (b) The display result of location
Figure 14. Navigation architecture of a centralized system
127
These technologies need a powerful computer but the vehicle location. For example, the navigation
may have some problems. In the increase of time system instructs the vehicle to turn right in 150m,
and advance of technologies, these problems will but it must turn right in 100m. In order to reduce
be overcome and vehicles will have a function of the error condition, we use a computer vision
automatic driving. system to correct the error. The combination of
a computer vision system, an automatic vehicle
consist of automatic Vehicle control system and a navigation system has the
capability to achieve an automatic vehicle.
From the descriptions of previous sections, we
know that the automatic vehicle will be the trend Navigation System
in the future. An automatic vehicle consists of a
computer vision system, a navigation system and In previous sections, the navigation system con-
a vehicle control system. The relation of these sists of a location system and map information
systems is demonstrated in Figure 15. First, the system which was introduced in Chapter 13.
computer vision processes a series of the image, The navigation system plays a very import role
which is capture by a camera, and then returns in the automatic vehicle that instructs the vehicle
the process result to the vehicle control system. when the vehicle makes turns. If a vehicle is lack
Second, the vehicle control system controls the of the navigation system, it will drive randomly
vehicle according to receive an information from and causes it dangerously. That is, a navigation
other systems. It only has the computer vision and system is an instructor telling a location with a
vehicle control systems are not enough because small error. The development problem is how to
the vehicle is lack of an indicator for the use of design a system with high precise location and
arbitrarily drive. Based on this condition, the how to use other related methods to improve the
provision of a navigation system for an automatic decision of automatic driving.
vehicle is necessary and important. The naviga-
tion system and the computer vision system are Computer Vision System
related with each other, because if there is an
error of location in the navigation system, the A computer vision system consisting of a camera
computer vision system can be used to correct and image process observes environments and
Figure 15. The components of automatic vehicle
Figure 16. Computer vision system
128
Figure 18. (a) An original image. (b) A grayed

responds to the control system. Figure 16 shows
image. (c) Vehicle detection. (d) A result of de-
an example of a scenario. Firstly, the camera
tection
captures images converted to digital images, and
then the image process detects some targets. In
consequence, the result of detection is shown in
the original image.
The function of a computer vision system can
be divided into three types: the lane detection,
the vehicle detection and the obstacle detec-
tion. First, the lane detection helps an automatic
vehicle to detect a lane which the vehicle drives
on. Second, there are many vehicles driving on
a lane. If a system can’t detect other vehicles, it
will cause accidents. Based on this reason, vehicle
detection is necessary. Finally, a lot of objects
may be on lanes such as people, animals, rocks,
traffic light, etc. The obstacle detection helps the roads are marked for each lane. For this reason, how
automatic vehicle to detect these situations and to handle this exception must be addressed.
to respond to the control system immediately. In In the vehicle detection, the function of the
this subsection, we first introduce the lane and vehicle detection is to detect other vehicles in front
vehicle detections. of the detecting vehicle. The vehicle detection must
In the lane detection, Jung et al., (2005) intro- consider some impact factors such as the distance
duced a lane detection technology. The procedures between vehicles, the speed of vehicle, etc. In this
include edge detection and boundary detection. subsection, we only consider the vehicle detection
In the edge detection, the edge distribution func- without other factors.
tion and Hough transform are adopted to detect Broggi et al., (1999) introduced the vehicle de-
edges. In the boundary detection, a linear model tection as indicated in Figure 18. First, it grays the
is adopted to detect boundaries. Figure 17(a) is image and then uses the symmetry detection that
an original image. The image shown in Figure detects a symmetry part of a vehicle for detecting
17(b) is grayed and then uses boundary detection vehicles. The symmetry detection uses the edge
to find the rough sketch of lanes. Finally, the lanes detection technology, and then uses the results of
are drawn in the original image as demonstrated vertical and horizontal edge detections to find an
in Figure 17(c). outline of a vehicle. It then uses the bounding box
In most related studies, they can detect the detection, detecting the left and right box on the
marked lanes; however, in the real roads not all of top or foot, to increase the result of detection. If a
Figure 17.
129
Figure 19. Structure of the vehicle control

reFerences
Brakatsoulas, S., Pfoser, D., Salas, R., & Wenk, C.

(2005). On Map-Matching Vehicle Tracking Data.
In The 31st VLDB Conference, (pp. 853 – 864).
Broggi, A., Bertozzi, M., Fascioli, A., Guarino,
C., Bianco, L., & Piazzi, A. (1999). The Argo Au-
tonomous Vehicle’s Vision And Control Systems.
International Journal of Intelligent Control and
Systems, 3, 409–441.
Chang, B.-J., Huang, B.-J., & Liang, Y.-H. (2008).
Wireless Sensor Network-based Adaptive Vehicle
box is not fined, it will use backtracking to correct
Navigation in Multihop-Relay WiMAX Networks.
results and repeatedly uses boundary detection to
The 22nd International Conference on Advanced
find vehicles.
Information Networking and Applications, (pp.
56-63).
Vehicle Control System
Farrell, J. A. (2008). Aided Navigation: GPS
The vehicle control system, an important compo- With High Rate Sensors (pp. 22-80). New York:
nent of the automatic vehicle, receives information Glencoe/McGraw-Hill School Pub Co.
from the computer vision and navigation systems,
Jung, C. R., & Kelber, C. R. (2005). Lane following
and then controls the vehicle. If the control system
and lane departure using a linear-parabolic model.
design has defects, it will cause accidents. Based
Image and Vision Computing, 23, 1192–1202.
on the reason, providing a stable and robust control
doi:10.1016/j.imavis.2005.07.018
system is required.
Thus, a control system must consider many Schiller, J., & Voisard, A. (2004). Location-Based
impact factors: error correction, speed of driv- Services. San Francisco: Morgan Kaufmann.
ing, etc. However, we only introduce a basic
architecture proposed in (Broggi et al., 1999), in
Systems. Boston: Artech House Publishers.
which the structure of the vehicle control system
is shown in Figure 19. A computer vision system Zhao, Y., & House, A. (1997). Vehicle Location
sends a result of image process to the vehicle and Navigation Systems: Intelligent Transporta-
control instructor, controlling the vehicle accord- tion Systems (pp. 43-78). Boston: Artech House
ing to vehicle information, image process and Publishers.
location. Finally, the vehicle obeys instructor’s
command to drive.
130
131
Chapter 9
Design and Implementation of
Vehicle Navigation Systems
Min-Xiou Chen
National Dong-Hwa University, Hualien, Taiwan, R.O.C.
absTracT
Vehicle Navigation System (VNS) is a complicated and integrated system. A reliable vehicle navigation
system should integrate the wireless communication technologies, positioning technologies, embedded
computer, geographic information database, and so on. The major purpose of the chapter is to help
understanding the architecture of vehicle navigation system. This chapter first introduces the system
requirements and system analysis, and show the system platform of vehicle navigation system. The system
platform can be divided into six components. There are the digital map database, positioning devices,
map-matching process, route planning process, route guidance process, human-machine interface, and
wireless communication interface. The design issues and system communication of these components
are detail illustrated in the chapter. Finally, the authors also present some vehicle navigation systems
proposed in the past few years, and show the difference of these systems. The aim of vehicle navigation
system is to guide the vehicle along the optimal path from the starting point to destination. A reliable
vehicle navigation system can reduce the traffic chaos in the city and improve the transportation delay.
In order to achieve reliable vehicle navigation system, the detail system requirements, system analysis,
and system architecture are shown in the chapter. Each component of vehicle navigation system is briefly
illustrated, and the system communication is also described. The authors also present the architecture of
the proposed vehicle navigation system, and show the difference of these systems. Therefore this chapter
helps understanding the architecture of vehicle navigation system.
DOI: 10.4018/978-1-60566-840-6.ch009
Design and Implementation of Vehicle Navigation Systems
inTroducTion system requirements
The aim of navigation system is to detect the posi- Let’s consider a scenario that a visitor is driving in
tion of the vehicle, track the vehicle and control a strange city, how can he know his position and
the movement of the vehicle from one place to plan the driving path to the targeted place based
another. The location techniques and the map are on his tourist map? Getting lost is a nightmare
two important components in the navigation sys- for each visitor. Thus, the system requirements
tem. In Age of Exploration, the shipman records for the tourist and the strange driver are shown
their position on the chart and pilots their courses as below (Zhao, 1997):
according the compass, sextant, and chronometer.
After World War II, the radar technique is imple- a. Shows the correct position of the current
mented to identify the range, altitude, direction, or vehicle or destination on the digital map;
speed of both moving and fixed objects. The radar b. Plans the shortest/fastest route from the cur-
technique is also involved in the modem naviga- rent position to the destination and show on
tion system in order to improve the correctness of the digital map;
the vehicle position, and sailing safety. In recent c. Guides the drivers along the planned
years, with the development of global positioning route;
technology, electronic technology, and wireless d. Tracks the vehicle on the digital map;
communication technology, the size of navigation e. Easy and safety to use;
system becomes smaller and can be carried on a f. Shows the real-time traffic informa-
bus, truck or car. Thus, the automotive navigation tion (emergency or congestion) on the
system had been proposed to guide vehicles in its monitor;
location using digital map display. Moreover, with g. Re-plan routers based on emergency infor-
rapid increment of automotives, the urban traffic mation or congestion information. Moreover,
becomes much more crowed and the traffic chaos more system requirements shown as below
is a serious problem in many cities. Therefore, had been proposed in the investigation results
how to design a vehicle navigation system in for the Taiwanese drivers .
order to reduce the traffic chaos and improve the h. Shows the warning message of the speed
transportation delay is a critical problem. In the traps or the over speed limit;
chapter, the detail architecture of vehicle naviga- i. Shows the park area and the gas station on
tion system is illustrated. the digital map;
j. Shows the shopping/restaurant information
on the digital map;
The design oF Vehicle k. Shows the tourist information on the digital
naVigaTion sysTems map;
l. Shows the Road-Side service;
The development process of an information system m. Reports the traffic accident; Summarize
should involve the system requirements, system these system requirements, the basic sys-
analysis, system design, system construction, tem requirements can be shown as the
system testing, and system maintenance tasks. following:
This chapter will focus on the system analysis, n. The function to show the location of the
system requirements, and system design of the current vehicle or destination on the digital
vehicle navigation system. map;
132
o. The functions to plan or re-plan the shortest/ the current position of the vehicle to the
fastest route; destination.
p. The functions to show the life e. Needs an interactive interface: The inter-
information; active interface involves the input devices
q. Friendly user interface. and output devices. The drivers can enter
the information into the system, or select
system analysis the information from the system by the input
devices, and see the results from the output
Analyzing the user requirements, the functions of devices.
the vehicle navigation system can be illustrated f. Needs a communication module: The
as following: communication module can receive the cur-
rent traffic information, such as emergency
a. Needs a digital map library: The digital information and congestion information,
map database is the key component for the from the traffic control center. The system
vehicle navigation system. The road map can re-plan a new route based on that current
should be digitalized into the digital library, traffic information. The vehicle navigation
before the system uses it. The location infor- system also can report the traffic information
mation of some government institutions and to the traffic control center.
some famous buildings should be digitalized
into the digital library, which can be denoted The implemenTaTion oF
as the landmarks. The system can obtain and Vehicle naVigaTion sysTems
show the relative position of each targeted
object on the map according to the location According to the system analysis, the system
information of these landmarks. architecture of the vehicle navigation system can
b. Needs the location modules: The location be shown as the Figure 1. The vehicle navigation
modules can be divided into the positioning system involves six components. There are the
module and map-matching module. The digital map database, positioning devices, map-
major function of the positioning module matching process, route planning process, route
is to get the position information from the guidance process, human-machine interface, and
position devices and obtain the coordinates wireless communication interface. The implemen-
of a vehicle on the surface of the Earth, or tation issues of these components are illustrated
the relative position of a vehicle on the city. as following section.
The aim of the map-matching module it to
provide drivers the correctly location infor- system platform
mation of the vehicle and show the precisely
position of the vehicle on the map. a. Digital map database: The digital map
c. Needs a route planning module: The route database is the key component for the ve-
planning module is a process that provides hicle navigation system. The information
the drivers to plan the route from the starting of the road map, the landmarks, such as the
point to the destination. government institutions and the famous
d. Needs a route guidance module: The route buildings, and the life information, such
guidance module is a process that guides as the hotel, park area, restaurant, gas sta-
the drives along the planned route from tion, and so on, also should be digitalized
and stored in the digital map database. The
133
Figure 1. System platform
basic representations for the digital road according to their importance. For example,
map are nodes, segments, and shape points. the nodes and shape points placed on the
A node can be represented a cross point or highways are in the upper layer, the nodes
an endpoint of a road. A segment can be and shape points placed on the country
represented the piece of road between two roads are in the lower layer, the landmark
nodes. A shape points is placed in the seg- points represents the airports, train stations,
ment, and is used to present the shape of the and harbors, are in the upper layer, and the
road. The location information of each node landmark points represents the park areas,
and each shape point, such as the latitude and restaurants, and gas stations, are in the lower
the longitude, are stored in the digital map layer. The vehicle navigation system can
database. The landmark points can be used retrieve these nodes and points according
to denote the landmarks and the life informa- the drivers requirement. For example, the
tion. The landmark points also contain the route planning process can only retrieve the
location information, such as the latitude, nodes on the highways and arterials, and
the longitude, and the address. These nodes, the map-matching process can retrieve the
shape points, and landmark points are very landmark points in the lowest layer when the
useful for the map-matching process. In vehicles is in the city. Moreover, the drivers
order to present the precisely road map, and can show the information of different layer
provide more useful information, a lot of the on the screen upon their requirement.
shape points and the landmark points will b. Location modules: The positioning devices
be created in the digital map database. The and the map-matching function are two vital
computation cost of vehicle location and components in the location modules. The
navigation will increase when the amount of major functions of the location modules are
these points increased. The hierarchical layer to get the coordinates of a vehicle on the sur-
is introduced to improve the computation face of the Earth from the position devices,
complex of vehicle location and navigation. obtain and show the precisely position of
These nodes, shape points and landmark a vehicle on the map. The most common
points can be classified into different layers used positioning devices in the positioning
134
devices can be classified into two groups. process can be classified into the single
The relative devices, such as transmission vehicle route planning and multi vehicle
pickup, wheel sensor and gyroscope, can routing planning. The single vehicle route
estimate the approximate position data, such planning can be solved by the shortest path
as the directions and travel distances. The algorithms, such as Dijkstra’s algorithm and
absolute devices, such as GPS and compass, Bellman-Ford algorithm. The single vehicle
can report the precisely position information. route planning is to determine the optimal
Both these devices can be used on the vehicle route under the time or fuel constraints. The
navigation system. The map-matching func- all-pairs shortest path algorithms, such as
tion receives the location information of the Floyd-Warshall algorithm and Johnson’s
vehicle from the positioning module, and algorithm, can be used to solve the multi
retrieves the neighbor landmarks from the vehicle routing planning problem. The multi
digital map library. Then, according to the vehicle routing planning is to determine the
location information of these landmarks, the optimal routes for all vehicles through the
map-matching function will obtain the place- minimal total tour length subject to the time
ment of the vehicle relative to landmarks, or or fuel constraints. The computation complex
roads, and show the result on the map. The of the route planning process is the critical
dead reckoning is the primitive technique problem. The hierarchical concept is intro-
used to determine the vehicle location rela- duced to improve the computation complex
tive to a reference point. When the starting of route planning. The route planning process
location and the previous displacements are can only retrieve the nodes on the highways
known, the vehicle position will be more and arterials, and the highways and arterials
reliable and accurate. The vehicle and dis- segments, and determine the optimal route.
tance traverse are widely used to estimate In addition, the divide and conquer method
the changes in the position of the vehicle can also be introduced to improve the com-
relative to the origin. The drivers also can plexity of the route planning process. The
input a query address into map-matching road map can be divided into several blocks,
function and see the precisely position on and the route planning process finds the path
the map. in each block. Finally, the route planning
c. Route planning process: The route plan- process assembles these block paths into
ning process is used to plan a path from the a complete route from the starting point to
starting point to destination. The starting the destination. The drivers can browse each
point can be the current position of the block paths on the user interface. Moreover,
vehicle, or entered by the drivers from the the route planning process should provide the
user interface. The drivers also can select re-planning function, when the emergency
the landmark from the digital library and event or congestion event occurs.
set as the starting point or destination point, d. Route guidance process: The aim of the
or input the destination information from route guidance process is to guide the drives
the user interface. Then, the route planning along the planned path from the current po-
module will retrieve the road map from the sition of the vehicle to the destination. The
digital library and plan the path from the start- route guidance process should retrieve all
ing point to destination. The planned route the road information involved in the planned
can be listed on the screen, and used by the route from the digital library, and list that
route guidance module. The route planning information, such as the road names, travel
135
distances, turns, and landmarks, on the output or touch screen, and enter the command in
devices. The drives can read that informa- the touch panels or touch screen. The driv-
tion before the trip. The route guidance ers also can send the voice commands into
process also can present that information the system by the microphone, and hear the
to the drivers in real time while en route. announce message from the speaker. A good
The route guidance process should retrieve design of the human-machine interface can
the current position of the vehicle from the improve the driving safety. The design prin-
location modules, and presents the proper ciples for a good human-machine interface
guidance messages, such as turn-by-turn were proposed in (Ligs et al., 1995).
instructions, to the output devices. The route f. Wireless communication interface: The
guidance process continuously retrieves wireless communication interface is an op-
the current position and direction of the tional component for the vehicle navigation
vehicle from location the modules, continu- system. The vehicle navigation system can
ously compares the position and direction provide route planning and route guidance
with the planned route, and shows the turn without the wireless communication inter-
message or some travel messages on the face. However, the wireless communication
user interface or provide a series of voice interface can improve the driving quality.
announcements to warn the drivers. When The wireless communication interface re-
the route guidance process finds the driv- ceives the current traffic information, such
ers off the planned route, a series of voice as emergency information and congestion
announcements should be provide to alert information, from the traffic control center.
the drivers driven back to the planned route. Once these messages sent, the vehicle navi-
The re-planning function provided by the gation system may alert to the driver. Then,
route planning process also can be invoked the system can re-plan a new route based on
to determine a proper path from the current that real time traffic information. The vehicle
position and direction to planned route, and navigation system also can report the traffic
the route guidance process guide the drives information to the traffic control center. The
back to the original route. Moreover, the re- traffic control center can get more precisely
planning function also can determine a new traffic information, determine the proper
route from the current position and direction traffic control policy, announce that informa-
to destination. When the emergency event or tion to all vehicles, and improve the traffic
congestion event occurs, the route guidance quality. Moreover, the traffic information
process should alert the drivers to plan the also can be exchanged between the vehicles
new route. without the traffic control center. This kind
e. Human-machine interface: The human- of network is referred as Vehicular ad hoc
machine interface provides a comfortable network (VANET), which a kind of wireless
and effective human use interface between ad-hoc network. The VANET includes some
the users and vehicle navigation system. The special characters, such as an open peer-to-
drivers control the system from the input peer network architecture, lack a central
devices, and read the information from the instance, each node is willing to forward
output devices. The input and output devices data for other nodes, self-configuration and
can be classified into the visual display based self-maintenance capabilities.
interfaces and voice based interfaces. The
drivers browse the digital map from the LCD
136
system communication advance
The system platform and system communication ADVANCE (Advanced Driver and Vehicle Advi-
of the vehicle navigation system is shown in the sory Navigation ConcEpt) (Ligs et al., 1995; Boyce
Figure 1. The digital map database is the kernel et al., 1994; Boyce 2002) is a project promoted
component, which provides the information of by the industrial, government and academic of
roads and landmarks to the map-matching process, Illinois, USA in July 1991. The project test was
route planning process, route guidance process, completed in the end of 1996. The ADVANCE
and human-machine interface. The digital map architecture is shown in Figure 2 (Zhao, 1997).
database can update the real time traffic data re- The Traffic related function is placed on the traf-
ceived from the wireless communication interface. fic information center, and can provide history
The map-matching process retrieves the roads traffic information and static traffic profiles. The
and the landmarks from the digital map database. dynamic traffic information reported from the
According that information and the vehicle in- mobile navigation assistant (MNA) are aggre-
formation provided by the position devices, the gated by the traffic related function. The traffic
map-matching process provides the precisely loca- information center (TIC) has many computers
tion on the map to the route guidance process and and devices to monitor the road network traffic,
shows on the human-machine interface. The route and report the traffic information to the MNA
planning process retrieves the roads, landmarks through the communication network. Both the
and real time traffic information from the digital TIC and MNA have the digital map database used
map database, and receives the real time traffic to show the road map. The MNA is carried by the
information from the wireless communication vehicle, and its architecture is shown in Figure
interface, plans a path from the starting point to 3 (Zhao, 1997).
destination, and passes the path to the human- The navigation computer is the key component
machine interface and route guidance process. The of MNA. The navigation computer gets the loca-
route guidance process retrieves current position tion information from the position devices, and
from the location modules, compares the position retrieves the map information from the digital
with the road information and real time traffic map database, which is a CD, and is read from
information in the planned route retrieved from the CD-ROM. The map-matching function, route
the digital map database, and announces proper planning function and route guidance function
turn-by-turn message to the drivers driven along are provided in the navigation computer. The
the planned path. The drivers also can browse navigation computer shows the road map on the
the map from the digital map database, and use touch screen, and announces the voice message
the route planning process to re-plan the route. from the speaker. The navigation computer also
The drivers can use the human-machine interface reports the traffic data which receives from the
to report their known information to the traffic position devices to the TIC, and receives the real
control center. time traffic information from the TIC through RF
modem. Due to the computation complexity, the
route planning function in ADVANCE only can
case sTudy consider the highway. The tourists are suitable for
the planned route obtained by the ADVANCE, but
Some vehicle navigation systems had been pro- the familiar drivers are not suitable.
posed in the past few years. Some of these imple-
ment architecture are illustrated in the section.
137
Figure 2. ADVANCE architecture
Figure 3. Mobile navigation assistant architecture
papago into four components. The map display compo-

nent is the digital map database, the information
PaPaGo is mobile navigation software imple- retrieval component used to search the object
mented by the Maction Technologies, Inc. in from the map component, and the route planning
Taiwan (PaPaGo SDK). The platform of PaPaGo is component is to plan the path from the starting
WinCE, and the implement environment is Visual point to destination. The route guidance function
C++. PaPaGo can be installed on the Pocket PC, and human-machine interface are integrated in the
Smartphone, and PDA. PaPaGo can be divided map display component. The position component
138
Figure 4. VETRAC architecture

of PaPaGo is GPS. Thus, the signal quality of GPS
is very important for PaPaGo.
VeTrac
VETRAC (Thangavelul el al., 2007), as shown in

Figure 4 (Thangavelul el al., 2007), is a vehicle
tracking and navigation project based on the
VANET. VETRAC does use the WiFi but not GPS
to track the vehicles. VETRAC deploy some WiFi
access point in the field, which can be used to re-
ceive the data sent from the mobile hosts, and also
can be used to monitor the traffic flow and track
the vehicles. The mobile computer with the client
control panel and display map system is installed
on the vehicle and uses the Mobile IP. The vehicle
will report the traffic information to the vehicle
tracking system and send some requirements to
the navigation server through AP or VANET. As
shown in Figure 5 (Thangavelul el al., 2007), the
navigation server is the core component of VE-
When new vehicle enters, the vehicle tracking
TRAC, and processes the traffic information sent
system will automatically report the positioning
from the vehicle tracking system, and response
server of the navigation server. The drivers see
the user requirements. The navigation server also
the location information from the display map
reports the location information to each vehicle,
system of client control panel.
and stores the traffic information, digital road map,
and vehicle locations into its database. Four servers
dmrg
are implemented in the navigation server.
The connection server is used to manage
Dual Mode Route Guidance System (DMRG)
the connections and the registries between the
(Muffat el al., 1991; Kuhne el al. 1995) is a
navigation server and the connection manager
vehicle navigation system promoted by the car
of the client control panels. The aim of the loca-
industries in the Europe. The system architecture
tion server is to response the location query sent
of DMRG is shown in Figure 6. DMRG integrates
from the location manager of the client control
the individual route guidance systems, and col-
panels. The traffic service can announce the traf-
lective route guidance systems. The individual
fic information retrieved from the client location
route guidance systems only plan a single vehicle
database to the traffic information manager of
route from the starting point to destination. The
the client control panels. The positioning server
collective route guidance systems placed on the
of the navigation server receives the traffic in-
traffic control center, can collect all the traffic
formation sent from the vehicle tracking system,
information, and provide the better planned router
and updates that information stored in the client
for each vehicle. The individual route guidance
location database.
systems use the on-board computer to plan the
The vehicle tracking system is used to ag-
path according to individual criteria (shortest
gregate the traffic information sent from the AP.
139
Figure 5. VETRAC software architecture
Figure 6. DMRG system architecture
way, quickest, simplest, etc). Thus, the several congestion due to the case. In contrast, the traf-
drivers use the same routing algorithm with the fic control center has each vehicle’s location and
same criteria and destination will have the same destination, the collective route guidance systems
planned path to follow. Some roads may become can plan different path for the vehicle with the
140
same destination, and the traffic congestion can Boyce, D. E., Kirson, A. M., & Schofer, J. L.
be reduced. The system architecture of DMRG is (1994). ADVANCE-The Illinois Dynamic Naviga-
very similar to ADVANCE. The major difference tion and Route Guidance Demonstration Program.
is the route guidance function. ADVANCE only Advance Technology for Road Transport, IVHS
considers the individual route guidance systems, and ATT.
but the DMRG integrates the individual route
Frost & Sullivan, (2005). Strategic Analysis of the
guidance systems, and collective route guidance
Taiwan Telematics and Infotainment Market.
systems.
Green, P., Levison, W., Paelke, G., & Serafin,
C. (1995). Preliminary Human Factors Design
conclusion Guidenlines for Driver Information Systems. Tech-
nical Report No. UMTRI-93-21, Transportation
In this chapter, we have shown the system ar- Research Institute, The University of Michigan,
chitecture and system communication of the Ann Arbor, MI.
vehicle navigation system. The user requirement
Kuhne, R. D., & Langbein-Euchner, K. (1995).
is described in the first. After system analysis,
Calculation of travel time savings by dual mode
the modules of the vehicle navigation system
route guidance for the South corridor in the Stut-
are proposed, and the functions of each module
tgart test field. Vehicle Navigation and Information
are briefly illustrated. The general operations of
Systems Conference.
the system communication are also described in
the following section. At last, we elaborate some Lee, J., Forlizzi, J., & Hudson, S. E. (2005).
architectures of the proposed navigation system, Studying the effectiveness of MOVE: a contex-
and describe the difference of these systems. In tually optimized in-vehicle navigation system.
recent years, human factors have been included In Conference on Human Factors in Computing
into navigation systems in order to develop the Systems.
intelligent transportation systems (Woodrow et
Ligs, J. F., & Bowcott, S. (1995). ADVANCE:
al., 1998; Lee et al., 2005), and to improve the
Initial Deployment. Annual Meeting of ITS
safety of drivers (Baldwin, 2002; Transportation
America.
Research Board, 2004). Summarizations these
system architectures, the navigation systems still Muffat, M., Green, P., & Crosnier, S. (1991).
have a large scope for development. European cooperation on dual mode route
guidance—Perspectives for advanced research
partners. Vehicle Navigation and Information Sys-
reFerences tems Conference. PaPaGO SDK. (n.d.). Retrieved
from http://www.papagosdk.com/
Baldwin, C. L. (2002). Designing in-vehicle
technologies for older drivers: application of Thangavelul, A., & Bhuvaneswari, K. Kumar’, K.,
sensory-cognitive interaction theory. Theo- SenthilKumarl, K., & Sivanandam, S.N., (2007).
retical Issues in Ergonomics Science, 3(4). Location Identification and Vehicle Tracking us-
doi:10.1080/1463922021000009029 ing VANET (VETRAC). In IEEE International
Conference on Signal Processing, Communica-
Boyce, D. (2002). A Memoir of the ADVANCE tions and Networking 2007.
Project. Journal of Intelligent Transportation
Systems, 7(2).
141
Transportation Research Board. (2004). Transpor- Woodrow, B., & Thomas, A. (1998). Dingus Hu-
tation In An Aging Society: A Decade of Experi- man Factors in Intelligent Transportation Systems.
ence. Transportation Research Board. Mahwah, NJ: Lawrence Erlbaum Associates.
Systems. Boston: Artech House.
142
Section 4
144
Chapter 10
Vehicular Metropolitan Area
Network Systems Architecture:
The WiMAX Network Reference Model
Cheng Hsuan Cho
National Chung Cheng University, Taiwan, R.O.C.
Jen-Yi Pan
National Chung Cheng University, Taiwan, R.O.C.
absTracT
The WiMAX NWG develops a network reference model to serve as an architecture framework for WiMAX
deployments and to ensure interoperability among various WiMAX equipment and operators. The net-
work reference model envisions unified network architecture for supporting fixed, nomadic, and mobile
deployments and is based on an IP service model. The authors introduce WiMAX network architecture,
WiMAX network entry, mobility management, QoS functional elements, core network planning and
accounting architecture in this section. However, all of them are significant in deploying WiMAX core
network. The operator tries to reach the goals including system performance, reliability, and so on.
On the other hand, the WiMAX operator should consider and balance such many variables in order to
achieve a better situation.
inTroducTion entities, and some important basic elements like

Network Service Provider (NSP) and Network
Based on WiMAX Forum Network Architecture Access Provider (NAP).
definition (Stage 2: Architecture Tenets, Reference The functional entities contains Subscriber Sta-
Model and Reference Points (Juo et al., 2008)) (Stage tion (SS)/Mobile station (MS), ASN(Access Service
3: Detailed Protocols and Procedures (WiChorus, Network) and CSN(Connectivity Service Network).
2009)), the WiMAX network reference model in- As in Fig 10.1, the network reference model depicts
cludes the logical functional entities and reference telecommunication’s operations. Access Service
points. Figure 1 shows the WiMAX network refer- Network is managed by NAP (Network Access
ence model containing reference points, functional Provider). CSN is managed by NSP (Network
DOI: 10.4018/978-1-60566-840-6.ch010
Vehicular Metropolitan Area Network Systems Architecture
Figure 1. WiMAX network reference model
Service Provider). CSN provides IP connectivity, contain reference points and functional entities
links out with the internet and combines with its in ASN. ASN-GW not only manages IP Data
Home-ASP (Application Service Provider). Basi- Forwarding in ASN, but also links with the other
cally, the overall architecture can be divided into function entities inside or outside the ASN-GW.
three parts. The first part (left position in Figure It tunnels data and packets to the suitable BS.
10.1) is MS/SS equipments, and it’s also called ASN-GW Control plane handles all of the radio-
the end users. The second part (center position independent control and includes authorization,
in Figure 10.1) is ASN. ASN supports wireless authentication, and accounting (AAA), context
communication and air-interface to MS/SS. The management, profile management, service flow
third part is CSN where inner management servers authorization, paging, radio resource management,
locate (such as Home Agent, Location Register, and handover. Data plane feature set includes
AAA server and so on). mapping radio bearer to the IP network, packet
Based on Figure 1, we give an introduction for inspection, tunneling, admission control, policing,
these functional entities and reference points: QoS and data forwarding.
ASN(Access Service Network): It links with ASP(Application Service Provider): ASP
WiMAX Customer Premises Equipment (CPE) provides basic network service and manages
e.g. MS, SS. Access Service Network (ASN) applications with IP-network. An application
functions support WiMAX Layer2 connectivity, service provider (ASP) is a business that provides
AAA messages forwarding, WiMAX NSP selec- computer-based services to customers over a net-
tion, Layer-3 link tunneling (BS-SS) and Radio work. Software offered using an ASP model is also
Resource Management (RRM). The operations sometimes called On-demand software.
can be seen in the ASN-GW function. CSN(Connectivity Service Network): The
ASN-GW(ASN Gateway): As in Figure 2, CSN contains several functional elements and
we can observe the ASN-GW functions which tasks for supporting IP-Network connectivity to
145
Figure 2. ASN-GW reference model
CPE. The following lists the major tasks in CSN example as Figure 10.1 shown, the NAP function
network: contains two ASN networks.
HA (Home Agent): HA supports mobility
1. MS/SS IP address assignment. management for MIP registration, de-registration,
2. Internet access. and packet tunneling in CSN, and it helps CSN
3. AAA Server management. maintain the IP address table or location area,
4. Tunnel supported (between CSN and SS). e.g. it records MS location where MS attached
5. Accounting /pricing service. the serving BS. So CSN can use this information
6. Users roaming across ASNs/ CSNs. to tunnel packets to MS/SS from corresponding
7. Mobility management between MS and node (CN). In the network reference model, HA
ASN. is located in CSN.
FA (Foreign Agent): Foreign agent is a router
NAP (Network Access Provider): ASP serving as a mobility router for a mobile node
provides basic network service and manages ap- (The definition is also in IETF-RFC-2002). A
plications with IP network. Moreover, it provides foreign agent works together with home agent to
radio link resource to one or more NSPs. One NAP support Internet traffic forwarding. For the MS/
can also contain one or more ASN network. For SS, it connects to the Internet from any location
other than its home network. The home agent
146
Table 1. Reference points R1~R8
Links Operation and Task

R1 MS -- ASN Implementation as the same as IEEE 802.16e-2005 definition
R2 MS -- CSN Authentication, authorization for MS, IP host configuration and mobility management
Supporting AAA, policy enforcement, and mobility - management capabilities (including tunneling and
R3 ASN -- CSN
forwarding)
R4 ASN-GW -- ASN-GW Used for MS mobility across ASNs
R5 CSN -- CSN For internetworking between home and visited network
R6 BS -- ASN-GW Implementation of intra-ASN tunnels and used for control plane signaling
R7 In ASN-GW Coordination between data and control plane in ASN-GW
R8 BS – BS Handover control, ensuring fast and seamless handover
tunnels packets intended for the mobile node to functional entities. We conclude these and show
a care-of address, which is either the IP address their work as in Table 1.
for the foreign agent, or an IP address acquired
through some external equipment or server, such
as Dynamic Host Configuration Protocol (DHCP) wimaX neTwork enTry
server. However, the foreign agent de-capsulates
packets and delivers them to the mobile node. The original network entry procedure is defined
(Detailed MIP operations are described in RFC- in WiMAX Forum stage2/3. The definitions are
3344-IP Mobility Support for IPv4 (Mustafa also described in 802.16e. However, the operation
Ergen, 2009)). can be separated by the following operations as
LR (Location Register): Location Registers shown in Figure 3.
are database functions storing information typi-
cally used in the routing of signaling information. nd&s (network discovery
Typical examples include the HLR (Home Loca- and selection)
tion Register) and VLR (Visitor Location Register)
used in telecommunication. 1. NAP discovery: NAP discovery is the first
BS (Base Station): BS serves Subscriber Sta- step of network entry. At first, an SS/MS
tions (point-to-multipoint), and provides SS with detects available NAP(s) by scanning and
first-mile (or last mile) access to public networks. decoding DL-MAP of ASN on detected
Besides, BS is directly connected to backbone net- channel to find NAP(s). SS/MS will range
works (e.g Ethernet). The complement definitions with the NAP if the NAP exists, then decode
of PHY/MAC layer are introduced in IEEE 802.16. DL-MAP, UL-MAP, DCD and UCD. (The
It can also support radio resource management: detailed operation is defined in IEEE 802.16
call admission control, packet scheduling, power d/e)
control and handover control. Base station has the 2. NSP discovery: The NAP may support more
local control of the network and gets assistance than one NSP. In Figure 4, if NSP Identifier
from ASN-GW for some features and implements Flag LSB in Base station ID equals to 1, it
the decision of the gateway for others. means there are more than one NSP in NAP.
RP (Reference Point): In NRM, Reference And also, MS acquires with NSP ID list.
point 1~8 are reference points between different After this process, MS must discover each
147
Figure 3. WiMAX network entry

MS indicates its NSP selection by attaching
to an ASN associated with the selected NSP,
and by providing its identity and home NSP
domain in form of NAI (network access
identifier).
authentication
This protocol implements 802.16e security with

IETF EAP framework. The protocol is used for
service flow authorization, mobility management
and policy control. AAA framework is based on
pull model in which supplicant sends a request to
ASN, and then ASN forwards it to AAA server.
The AAA returns with appropriate response
to ASN which sets up the service and informs
the MS. The elements are depicted in Figure 5.
NSP until all of these NSP are founded. On The operation applies PKMv1 and PKMv2 by
the contrary, if NSP identifier Flag LSB is Extensible Authentication Protocol (EAP) to
equal to 0, MS will knows there is only one authenticate with user. PKMv1 is authentication
NSP service behind the NAP. for device. In contrast with PKMv1, PKMv2 is
3. NSP enumeration and selection: After NSP authentication for user and Device between CSN
discovery, MS produces a list of available and MS. In the authentication flow, user Authenti-
NSPs as discovered through NSP discovery cation is optional, while device Authentication is
in the available NAPs, MS uses the list select Mandatory. When the authentication procedures
NSP by manual or automatic selection. just only include Device Authentication or User
4. ASN attachment based on NSP selection: Authentication, Single EAP can support the
Following a decision to select an NSP, an procedure. But when user and device both need
Figure 4. Base station ID format (Juo et al., 2008)
148
Figure 5. WiMAX user authentication protocol
authenticated, double EAP is needed, and each data Transfer

uses different AAA server (e.g. AAA servers is
located in different CSN). At this stage, MS/SS is able to use WiMAX core
network. The user can use the data link to browse
configuration setup web-sites and transfer data. But when the user
moving to another BS’s coverage, not only the
The procedure of Configuration Setup is after attached WiMAX ASN-gateway will change, but
Authentication procedure. In order to support dif- also the IP address will change. We will discuss
ferent network architecture (e.g. IPv4 or IPv6, net- the mobility issue in the later paragraph about
work with or without DHCP), many configuration moving between ASN/CSN scenarios.
processes are proposed: ProxyMIPv4 (PMIP4),
ClientMIPv4 (CMIP4), Proxy PMIPv6 (PMIP6)
and Client MIPv6 (CMIP6). The “Client” means wimaX mobiliTy managemenT
MS/SS support Mobile IP (MIPv4 / MIPv6) func-
tion. It doesn’t need other equipments forward or The WiMAX network supports two types of
tunnel packet. Proxy Mobile IP (MIPv4 / MIPv6) mobility: (1) ASN-anchored mobility and (2)
means Proxy server is located in ASN which helps CSN-anchored mobility. ASN-anchored mobility
MS/SS in charge with registration and signaling is also referred to as intra-ASN mobility, or mi-
MIP packets to HA. Therefore, although MS/ cromobility. In this case, the MS moves between
SS doesn’t have capability to supporting MIP, it two data paths while maintaining the same anchor
can still roams in the wireless network by using foreign agent at the northbound edge of the ASN
Proxy-MIP server. network. The handover in this case happens
across the R8 and/or R6 reference points. ASN-
anchored handover typically involves migration
of R6, with R8 used for transferring undelivered
packets after handover. It is also possible to keep
149
the layer three connection to the same BS (anchor Mobility and CSN Anchored Mobility function
BS) through the handover and have data traversal and operation:
from the anchor BS to the serving BS throughout
the session. asn anchored mobility
On the other hand, CSN-anchored mobility
is also referred to as inter-ASN mobility, or ASN anchored mobility takes place when MS
macromobility. In this case, the MS changes to a moves to a neighboring BS that connects to a
new anchor FA, this is called FA migration. The different ASN-GW, which may be within the
new FA and CSN exchange signaling messages to serving ASN or not. If the target BS belongs to a
establish data-forwarding paths. The handover in different ASN, the target ASN-GW will establish
this case happens across the R3 reference point, a R4 data path to the anchor ASN so that the MS
with tunneling over R4 to transfer undelivered can avoid data loss and the QoS could also be
packets. Figure 6 illustrates the various possible guaranteed. During the ASN anchored handoff
handover scenarios supported in WiMAX. procedure, the MS does not change its CoA.
Based on different mobility scenarios (e.g. ASN Anchored Mobility divides ASN function
ASN-anchored mobility, CSN-anchored mo- into preparation phase and Action phase. 802.16e
bility), we can see Intra-ASN mobility and working groups defined MOB MSHO-REQ (when
inter-ASN mobility between the WiMAX core the network initiated, we called MOB BSHO-
network. WiMAX End-to-End Network Systems REQ) message format for preparation phase and
Architecture Stage 3 shows ASN Anchored defined MOB HO-IND message format for action
phase, some of preparation’s steps is optional. It
Figure 6. Mobilityscenario with multiple ASN-GW
150
can keep these steps on preparation phase until anchored mobility the traffic comes from ASN-1.
action phase. On the contrary, if we did these ASN-2 is the new anchor ASN for MS.
steps in preparation phase, the action phase leaves Based on these mobility scenarios, we make a
out the operation time that can reduce a part of simple conclusion and comparisons in Table 2.
handover latency.
For ASN anchored mobility, traffic to MS ho Function network Transaction
comes from the anchor ASN. Figure 10.7 depicts
the data flow of ASN anchored mobility. ASN- The handover (HO) allows MSs to handover
1 is the serving and anchor ASN to MS. After between neighboring BSs while moving across
handover, the packets from are CSN delivered the corresponding coverage areas. Furthermore,
from ASN-1 to ASN-2 via a R4 data path. The the mechanism can be used by BSs to trigger a
anchor ASN to MS remains the ASN-1. HO in order to optimally balance the traffic load
of cells within a network.
csn anchored mobility The basic HO function transaction is shown
in Figure 9. HO function contains serving HO,
On the contrary, CSN anchored mobility involves relaying HO and target HO.
in anchor ASN relocation, which means anchor
ASN is changed after CSN anchored mobility. 1. At first, the serving HO function initiates an
Figure 8 describes the data flow after CSN an- HO network transaction by sending HO_Req.
chored mobility. The packets directly send from There can be only one serving HO function
CSN via the new R3 link to ASN-2, unlike in ASN for any given HO network transaction. After
Figure 7. ASN anchored mobility scenario
151
Figure 8. CSN anchored mobility scenario
Table 2. Comparison with ASN- anchored mobility and CSN- anchored mobility
Mobility type ASN- Anchored Mobility CSN- Anchored Mobility

Handoff latency Faster – handoff process without through FA Slow– handoff process through FA
CSN changes FA. MS re-registers with CSN and it’s
ASN-GW’s FA does not change, and the original FA tun-
FA change or not HA with MIP registration procedure. Then, new FA will
nels packets to new BS.
forward packets.
When MS move to another ASN, HA doesn’t MS will change its CoA on HA’s mapping table when
From HA viewpoint
knows that. it re-registers.
Backhaul Link Ef-
-- Better
ficiency
receiving HO IND from MS, serving HO 3. Serving HO Function and target HO func-
function the serving HO function confirms tions communicate either directly or with
HO to only one target HO function by send- assistance of one or more relaying HO
ing HO Confirm messages. function. If the serving and target HO func-
2. The target HO Function responds to the HO tions cannot communicate directly for any
network transaction with HO_Rsp. There reason, the relaying HO function takes care
can be one or more target HO functions for of delivering the relevant information to the
an HO network transaction. corresponding target HO functions. A single
HO primitive (e.g. HO_req) that is sent
152
Figure 9. HO function transaction
from the serving HO functions may contain the standard are defined to ensure the end-to-end
information relevant for several target HO QoS:
functions. In this case several behavioral
policies might be applied. 1. Pre-provisioning of the service flow
parameters.
Qos FuncTional elemenTs 2. Signaling function to dynamically estab-
lishing the QoS enabled service and traffic
Under the IEEE 802.16 specification, a SS could parameters.
be associated with a number of service flows 3. Utilization of MAC scheduling and QoS
characterized by QoS parameters, and QoS frame- traffic parameter for uplink service flow
work for the air interface. The specification also 4. Utilization of QoS traffic parameter for
defines combined scheduling scheme, Resource downlink service flow.
allocation and admission control. Besides, the 5. Grouping of service flow properties into
WiMAX forum also describes QoS framework for name Service Classes.
WiMAX Network service, as shown in Figure 10.
The WiMAX Forum has specified a framework Based on Figure 10, we give an introduction
for Service Management and QoS. Service Flow for these functional entities:
(SFA) and Service Flow Manager (SFM) are the
entities that act as policy decision and enforce- 1. SFA (service flow authorization): Is re-
ment points respectively for Service Management sponsible for evaluating any service request
and QoS. The following figure depicts the policy against the subscriber’s QoS profile. SFA
framework for WiMAX networks: logical entities in the ASN. In case the user
As service flow provide a particular QoS QoS profile is downloaded from the AAA
according to the QoS parameter set defined for into the SFA at network entry phase, the SFA
that service flow. The following mechanisms of is responsible for evaluating any service
request against user QoS profile. The SFA
153
Figure 10. QoS functional elements (Juo et al., 2008)
also performs ASN-level policy enforcement their respective databases. The databases
using a local database and an associated local include general policy rules as well as
policy function (LPF). The LPF can also be application-dependant policy rules.
used to enforce admission control based on 4. AF (application function, AF): Hosts the
available resources. service logic and communicates the appli-
2. SFM (service flow management, SFM): Is cation level session information to PCRF
responsible for the creation, admission, acti- such as classifiers identifying service flows,
vation, modification, and deletion of 802.16 on which policy control and differentiated
service flows. It consists of an Admission charging is required.
Control (AC) function and the associated 5. AAA Server: Holds the subscriber’s QoS
local resource information. profile and the associated policy rules per
3. PF (policy function, PF): Resides in both subscriber. At first, when the user enter
home and the visited network, comprising WiMAX network,AAAserver will download
154
users QoS parameters for Authentication and businesses, OPEX may also include the cost of
Authorization. After that, AAA server can workers and facility expenses such as rent and
also release these QoS parameters to PF that utilities. In the WiMAX core network structure,
PF can decide how to handle the message OPEX are costs associated with the operation and
form AAA server. maintenance of an income producing property. It
includes accounting expenses, advertising, office
expenses, supplies, attorney fees, insurance and
wimaX core neTwork planning property management.
C. Matching data density requirements to base
We believe that planning and modeling are impor- station capacity (Akyildiz et al., 1999)
tant things for build a good network structure. In D. For WiMAX deployment scenarios, data
this section, we suggest some factors needed to density is an good metric for matching base station
consider in planning a telecommunication network capacity to market requirements. Demographic
- 1. Network deployment 2. Quality of Service information, including population, households,
management 3. Mobility management 4. Network and businesses per sq-km or per sq-mile, is read-
management. Based on these aspects, we’ll focus ily available from a variety of sources for most
on those managements and give more details: metropolitan areas. With this information and
1. Network deployment: Network deploy- the expected services to be offered along with
ment is the first consideration for telecommuni- the expected market penetration, data density
cation/wireless communication operators. The requirements are easily calculated. This 6-step
WiMAX operator needs to figure out the capital process is summarized in Figure 11.
expenditure and operation expenditure on their 2. QoS of network management: The 802.16
deployment scenario. standard provides some QoS designs in order to
A. CAPEX: The capital expenditure (CAPEX) achieve service quality demands, such as voice
includes the costs of wireless end device and and video quality. Also, ISPs have their own
WiMAX equipment (e.g. BS/SS/RS), backhaul QoS management schemes to provide different
link equipment, server and edge equipment. All fees for service items, which means the provided
of the costs when we set up our environment are solutions can be customized. Therefore, adequate
contained in CAPEX. For example, spectrum collocation of service levels and charges must be
license is included in CAPEX, many WiMAX considered so as to satisfy customers with fine
equipments (802.16e BS/SS) are based on li- revenue.
censed bands (2.5G/3.5G), but some are based 3. Mobility management: A user can move
on un-licensed bands. In licensed bands, we from the serving BS’s coverage to another
need to consider how much bands that we need target BS’s service coverage. In this scenario,
to cover and how much cost we need to pay for user’s move involves handover processes. In
these bands which are differ from continuous or fact, Handoff processes consist of three phases:
non-continuous bands. The first WiMAX-certified handoff decision, radio link transfer and channel
products will be operating in the licensed 3.5 GHz assignment (WiMAX Forum, 2007). Considering
frequency band, followed by systems for both the in handoff decision, several goals are important
2.5 GHz licensed band as well as the 5.8 GHz while designing and evaluating mobility manage-
license-exempt band. The important thing is to ment procedures:
add all of this payment in our CAPEX.
B. OPEX: An OPEX is an on-going cost A. Small signal overhead over the air
for running a product. For larger systems like interface
155
Figure 11. Determining market driven capacity requirements (Akyildiz et al., 1999)
Table 3. Factors of handoff decision
Factors Details of factor

Network Bandwidth Available bandwidth of a network
provider
Coverage Coverage of a network
Related service Academic/ enterprise/ commercial
Connections Support numbers of user
Mobility Support mobility management or not
Network structure WLAN/3G/3.5G/WiMAX
User Application The application running on mobile node(MN)
Power Power consumption of MN
Price Offers service’s fee to ISP or network provider
RSS Received Signal Strength from base stations
Bandwidth The minimum bandwidth that user needed
Packet loss Packet loss of the connecting network
B. Low handover latency wimaX accounTing

C. Minimum packet losses during handover archiTecTure
D. Least handover failure
E. Efficient use of network resources How to account and price in Wimax network?
802.16e working groups defined AAA- Authen-
For the Network provider and user provider, tication, Authorization and Accounting, and those
those are many factors that we can consider in. are a standard for security pricing system. About
However, we conclude from these observations. the Authentication, we are already discussed
Table 3 shows the factors on handoff decision before. It implements security procedure with
(WiMAX Forum, 2007): IETF EAP framework. After that, WiMAX opera-
4. Network management: While the network tor accounts for MS/SS based on some account
is setup, the operation and management of WiMAX procedure.
network are required to handle by WiMAX opera- Accounting architecture is shown in Figure
tor. Those are the key points to maintain network 10.12. The figure shows network elements for
service. We also divide it into the following parts both postpaid and prepaid services. The figure
in Table 4. also shows the network elements for Hot-lining
support and negative volume count for ASN. A
156
Table 4. Factors of network management
Content Content analysis

Entity management BS/ SS/ Router/ Switch management, and so on.
Integration WiMAX network and backbone network or others
Integration
network service (e.q PSTN, 3G)
Includes network resource as IP assignment and account for usage
Inventory management
of bandwidth
Network Planning Coverage Ratio, radio resource and related network management
Network service Charge user’s account and deal with permission of user
Manage and price service level agreements (SLAs) for differentiated services
Billing service that use a simple upper bound for the effective bandwidth of the conforming
traffic as a proxy for resource usage.
Good running of customer relationship management give an
insight for firms to understand the market trend so that firms can
customer relationship management (CRM)
adapt the changing environment and anticipate the future needs
(e.g. web-site, customer service, maintenance)
It can be used to uncover hidden patterns in data that have been
Data mining
collected (e.g. user’s preference)
description of each entity is provided in the fol- accounting, QoS-Based accounting, and online
lowing sections. (prepaid) accounting. These include the delivery
NWG defined the Accounting Architecture of information for the purpose of billing (both
(Juo et al., 2008). As shown in Figure 10.12, it prepaid and post paid billing) and information
contains NAP, Visited NSP and home NSP. All that can be used to charge activity by both the
of the operation and data flows are based on RA- home NSP and visited NSP. These accounting
DUS. The architecture defined offline (post-paid) procedure are introduced as following:
Figure 12. WiMAX accounting architecture
157
offline (post-paid) accounting online (prepaid) accounting
We describe the off-line (post-paid) accounting This section describes the online (prepaid) bill-
procedures and the Usage Data Records (UDRs) in ing procedures in the WiMAX core network.
this part. The Serving ASN must merge radio spe- The prepaid packet data service allows a user to
cific parameters (number of bytes/packets transmit purchase packet data service in advance based on
at the BS) called Airlink Records with IP network volume or duration Account status is stored on a
into the Usage Data Records (UDR). After that, the prepaid server (PPS) that is located in the user’s
Serving ASN uses RADIUS accounting messages home network and accessed via the HAAA server.
to send UDR information to the home RADIUS To provide service to roaming prepaid users, the
Server (via the visited AAA server if the subscriber visited ASN or CSN needs to support the prepaid
is roaming). The Visited and Home RADIUS server service and the local and broker AAA servers need
should be support the Accounting-Request records to forward the new prepaid accounting attributes
as specified in (WiChorus, 2009) specification. transparently to and from the home AAA server.
Based on the UDRs, Home AAA server can use The HAAA server and the prepaid server could
the records manage and account for MS. be collocated or could be separate entities.
However, the architecture is a definition
Qos-based accounting for accounting procedure. Mostly, the operator
combines the online, offline, and QoS based ac-
The QoS-based accounting is based on user’s SLA. counting method in their service, and they use the
For the basic service, WiMAX BS supports the UDRs to charges for subscribers. For example, the
802.16 defined QoS classifications. But when user WiMAX operator needs to price for their subscrib-
bill payment for ensuring connection/bandwidth ers. For simply, offline and QoS accounting are
thresholds, the UDRs are also includes these param- pricing schemes for web surfing or downloading/
eters for discount if there is a mistake for wireless uploading of each subscriber. We denote it by
connection .for example, if the user chooses the application service fee. Conversely, most opera-
high bandwidth scheme, the Home AAA server tors charge the basic service fee in every month,
in CSN should support this pricing scenario for and come together with the service fee that we
bandwidth accounting. On the other hand, based on called basic service fee. However, by using two
different classifications, the QoS-based accounting -part tariff strategy to subscriber is not always
supports different applications. For example, when better. Depending on different strategy of WiMAX
we use the FTP, WiMAX operator may consider operator, the pricing scheme can be changed or
user’s FTP’s UDR and charge activity with the modified. Regarding the accounting methods, the
amount of FTP session’s packet. The scheme important thing is that WiMAX operator should
may also consider the time of user’s connection. forecast the profit of a broadband access system
For example, WiMAX operator focused on VoIP in MRT environment.
application and Video on demand (VOD) applica-
tion and charge with the time of these applications.
Based on these requirements, the Visited RADUIS, reFerences
Home RADIUS, ASN-gw and BS should support
the application’s accounting-Request records. In Akyildiz, I. F., Ho, J., & Wang, W. (1999). Mobility
this way, QoS-based accounting can be used in Management in Next-Generation Wireless Sys-
WiMAX core network. tems . Proceedings of the IEEE, 87(8), 1347–1384.
doi:10.1109/5.775420
158
IETF Network Working Group. (2008). RFC3344- WiMAX Forum. (2007). WiMAX End-to-End
IP Mobility Support for IPv4. Retrieved from Network Systems Architecture Stage 3: Detailed
http://www.faqs.org/ rfcs/rfc3344.html Protocols and Procedures, Release 1.1.0.
Juo, C. S., & Pan, J. Y. (2008). Software Agent WiMAX Forum. (2007). WiMAX End-to-End Net-
Framework for Dynamic Handoff Decision. In work Systems Architecture, (Stage 2: Architecture
APCC 2008, Akihabara, Tokyo, Japan. Ergen, M. Tenets, Reference Model and Reference Points),
(2009). The Access Service Network in WiMAX: Release 1.1.0.
The Role of ASN-GW.
WiMAX Forum. (n.d.). Wimax deployment con-
WiChorus, Inc. (2009). Retrieved from http:// sider for fixed wireless Access in 2.5GHz and
www.mustafaergen.com/asn_gateway.pdf 3.5GHz licensed bands.
159
160
Chapter 11
Interworking of IP Multimedia
Subsystem and Vehicular
Communication Gateway
Wei-Kuo Chiang
National Chung Cheng University, Chiaya, Taiwan, R.O.C.
An-Nie Ren
National Chung Cheng University, Chiaya, Taiwan, R.O.C.
Abstract
In recent years, more and more people dream of experiencing various IP-based multimedia application
services when they are driving through their car. However, those multimedia devices in the car may
use different communication protocols such as X.10, Havi, Jini, UPnP and SIP. In order to provide a
variety of IP-based multimedia services to those users in the car, the authors mainly investigate the is-
sue of interworking between IP Multimedia Subsystem (IMS) and telematics of the vehicular industry.
A service-integrated platform, Open Service Gateway Initiative Service Platform (OSGi SP), has been
proposed to act as a Residential Gateway (RGW) and to administer the communication between the
vehicular environment and Internet. Besides, a Home IMS Gateway (HIGA), which can be implemented
on a NGN RGW, has been developed by Home Gateway Initiative (HGI) since 2005 to collect the rel-
evant information of in-car users, devices and services and to manage the IMS sessions for the in-car
devices that do not support IMS functions. With these techniques, the users can enjoy their digital life
by interacting with the home/vehicular network from anywhere.
Introduction different communication protocols. Due to the lack

of interoperability between all devices on the car,
Vehicular network has emerged as a hot research a service-integrated platform is needed to provide
issue recently. More and more people want to ex- the characteristic of interoperability. On the respect
perience a variety of IP-based multimedia applica- of home network environment, it will encounter
tion services when they are driving through their this problem as well. With the wide spreading of
cars. However, those devices on the car may use the digital home products, Home Networking has
gradually been becoming a popular issue in recent
DOI: 10.4018/978-1-60566-840-6.ch011 years. The Home Networking is a way to manage all
Interworking of IP Multimedia Subsystem and Vehicular Communication Gateway
kinds of network devices or home electrical appli- the applications of remote accessing the home/
ances; it allows the users to operate and manage vehicular network, we will discuss how to sup-
theses networking electrical appliances through port SIP on OSGi platform, how to utilize SIP as
a unified interface or platform. These electrical an unity way for message control, and we will
appliances represent all kinds of facilities that further probe and discuss the mapping of control
have specified functions and could be controlled messages between SIP and UPnP in order to solve
or administered by using specified network media the interoperability problems among the devices
or technology, e.g., X.10 device, HAVi device, that adopt different communication protocols.
Jini device and UPnP device. One of the goals of Since the IP-based network technology has
the Home Networking is to control the in-house become more popular, the future communication
electrical appliances from the external network by network will move toward the service integration
using hand-held devices (such as mobile phone, of the all-IP network and use the IP Multimedia
PDA) to connect to the home gateway through Subsystem (IMS) as a core network to provide a
the IP network, and thus to control the electrical variety of multimedia services. Therefore, inte-
appliances inside the Home Networking (e.g., grating the IMS and the vehicular communication
lamps, air-conditioner, TV). Besides, if there hap- environment will be a developing trend for future
pens any status changing to the in-house devices, network service providers. A user could then
it (the networking system) could notify the user monitor the in-car devices or access the data in
immediately. In addition to the Home Networking the multimedia storage device through his mobile
technology, we could also treat a car/vehicle or any device in anytime and anywhere, this will bring
public transportation as a mobile home network the user a richer experience in vehicular com-
to integrate the home network and telematics of munication network.
the vehicle industry. In order to fulfill these goals,
there are yet some issues necessary to be discussed,
e.g., security, mobility, the interoperability among relaTed works
the communication protocols, etc.
In recent years, because the growth speed of sip (session initiation protocol)
the mobile devices and in-house network devices
becomes very fast and the communication proto- SIP (Session Initiation Protocol) (Rosen et al.,
cols and home network/vehicular communication 2002) is a communication protocol developed by
technologies have undergone diversified changes, the MMUSIC (Multiparty Multimedia Session
these have made the intercommunication among Control) task group of IETF (Internet Engineering
the-mentioned devices to be more difficult. For Task Force). SIP is an application layer protocol for
example, the communication protocols adopted session control and signaling control, it could be
by the in-house/in-car devices could probably used to initiate, modify and terminate sessions. The
be UPnP, IP and SIP; therefore, an Open Service SIP application range is very extensive, including
Gateway Initiative (OSGi) has been proposed to voice and video calls over Internet, video con-
construct a service-integrated platform among ferencing, presence service, event subscription/
the Home Networking, vehicular communication notification and instant messaging. In November
environment and the Internet. of year 2000, SIP has been accepted by (The Third
Because the SIP communication protocol Generation Partnership Project [3GPP], 2009) to
possesses the capabilities of security mechanism, become the protocol for conveying communica-
event notification, media streaming and mobility tion control _ignaling and has been applied in the
management among the terminals, so, concerning IMS (IP Multimedia Subsystem) infrastructure of
161
the Next Generation Network (NGN). This SIP SIP User Agent will contain both UAC and
protocol could be used to establish and administer UAS functions.
the sessions in real time among the users, which • SIP proxy: SIP Proxy is in charge of pass-
includes the application services such as text, ing the received request to another SIP
voice, picture, video and interactive gaming, component or SIP UA. When a SIP UA
etc. It could also invite other users to join in an sends a request, this message will not be
established communication session to provide directly sent to the SIP UA at the destina-
a multicast conference service. SIP is similar to tion terminal, it could be passed through
HTTP (Hyper Text Transmission Protocol), both multiple SIP proxies before the request
are text-based communication protocols; SIP URI message will be sent to the terminating SIP
(SIP Uniform Resource Identifier) address is ex- UA.
actly like an e-mail address, e.g., Alice@example. • Registrar server: Registrar Server is re-
com.In the following, we will simply introduce sponsible for processing the SIP UA regis-
the fundamental elements of SIP: tration, and it is used to manage the certain
specified service as well as to update the
• SIP user agents: User Agents are the location information for the SIP UA.
terminal equipments in the SIP network, • Redirect server: When the Redirect
which could be SIP phone or the SIP client Server receives a request from SIP UA or
software in a PC; it includes User Agent proxy, it will return with a Response of 3xx
Client (UAC) and User Agent Server to inform the SIP UA or the proxy that this
(UAS). UAC is responsible for generat- request should be directed to another SIP
ing requests, and UAS is responsible for component.
generating the corresponding responses
with respect to the received requests. Each The SIP session establishment flow chart is
shown in Figure 1. Firstly, the terminating UA
Figure 1. Architecture of SIP message flow
162
must register with the Registrar Server; the reg- provides the multimedia services such as video
istrar will then save the UA location information phone, conference, VoD, OMA service enablers,
into the Location Server. After that, the originating etc.
UA will send the request to the outgoing proxy; The IMS (3GPP TS 23.228, 2008; 3GPP TS
the proxy will make an enquiry against the DNS 23.218, 2008) (see Figure 2) formulated by 3GPP
server for the incoming proxy address of the adopts SIP protocol as the communication basis,
terminating UA, and then pass this request to the it is a multimedia communication platform that
incoming proxy. The incoming proxy of terminat- provides functions of session management, secu-
ing UA will then interrogate the Location Server rity, mobility, QoS and charging. IMS is widely
for the terminating UA address, and pass this accepted as a network framework core that could
request to the terminating UA. When a session has realize the Fixed Mobile Convergence (FMC) and
been established, the caller and the callee could Triple Play Services and provide strong service
communicate by using RTP (Real-time Transport supports.
Protocol) or other transmission protocols. The Call Session Control Function (CSCF) of
IMS is the core component of IMS. Its purpose is
ims (ip multimedia subsystem) to process the signaling control for session set-up
between the users or between the user and the
The Next Generation Network (NGN) has in- server; it includes completing registration, basic
tegrated all types of heterogeneous networks call control, SIP signal route control, service
to provide the diverse, novel and personalized trigger, etc. According to functional difference in
multimedia application services. By observing IMS, CSCF could be classified into Proxy Call
the past telecom network development, we could Session Control Function (P-CSCF), Interrogat-
see that, from 2G/2.5G to 3G/B3G and even the ing Call Session Control Function (I-CSCF) and
3.5G/4G, the service technology development Serving Call Session Control Function (S-CSCF).
has been evolving gradually from the traditional The role of P-CSCF is just like a proxy server
voice communication service into the nowadays that is responsible for forwarding the SIP mes-
all-IP network framework that uses the IMS (IP sage in IMS. P-CSCF also acts as an User Agent
Multimedia Subsystem) as its core; the system (UA) in abnormal condition; I-CSCF is mainly in
Figure 2. 3GPP IP multimedia subsystem
163
Figure 3. S-CSCF service triggering in IMS
charge of interrogating Home Subscriber Server added-value service subscribed by the user within
(HSS) and selecting a S-CSCF for the user, then the service call control component (S-CSCF). If
forwarding the SIP message to the S-CSCF; the they are conformable, then the application server
role S-CSCF plays is a registrar, it also controls (AS) should be added into SIP call control signal-
the execution of the communication (call) and ing path; AS will monitor the call session set-up
service triggering. status and decide whether to execute the added-
Home Subscriber Server (HSS) is the main value service subscribed by the user.
database of 3G telecom system. It is in charge of
storing the criteria of service triggering for the osgi (open service gateway initiative)
user subscribed services. Application Server (AS)
is responsible for executing all kinds of service The full name of OSGi is (Open Service Gateway
logics, such as call forwarding, call waiting, voice Initiative, 2009). The generally mentioned OSGi
mail, etc. In addition, AS also provides SIP-based represents two meanings; it could represent the
service such as the services formulated by (Open OSGi Alliance organization and could also be the
Mobile Alliance, 2009) ; those include Multimedia service specification - OSGi service platform.
Message Service (MMS), Push to talk over Cellu- OSGi service platform is mainly developed bas-
lar (PoC), mobile location service, etc. It is able to ing on Java language; OSGi is a standardization
provide the user with more extensive multimedia organization founded by Sun, IBM, Ericsson and
services. And, S-CSCF could communicate with others in March, 1999. Its main goal is to provide
AS that supports SIP interface by going through a complete solution for point to point service de-
IP Multimedia Service Control (ISC) that adopts livery between remote service provider and local
SIP as the communication protocol. devices. Therefore, OSGi Alliance has defined a
The IMS service triggering flow has been de- open platform that allows the application program
fined in 3GPP TS 23.218 (see Figure 3). The initial and its value-added service of the remote service
requests sent by the user will be cross-compared to provider to be downloaded at any time to the
initial Filter Criteria (iFC) that corresponds to the Gateway nearby the user according to the user
164
Figure 4. OSGi framework layering
demands; and it could be automatically installed. module layer, life cycle layer, service layer, and
The gateway is a common device used to connect etc.; as shown in Figure 4.
the home/vehicular network to Internet. Bundle is an application program on an OSGi
Observing from the framework view, OSGi framework, it could be activated and executed by
could be decomposed into three componenets: the OSGi framework. Since it is an application
Framework, Bundle and Service. The framework program, to define the specified file format is re-
is constructed on Java virtual machine and the quired. An OSGi bundle is a Java Archive (JAR)
application program executed on the framework file; and the componenets included in each JAR
is called bundle; the interface service provided file are respectively Java class, Activator class,
by or required by each bundle is called a service. Manifest header and some resource files (such as
Bundle could be downloaded from the remote HTML web page or some graphic files); in which,
service provider and be automatically installed the Manifest header mainly contains the descrip-
on the framework; after installing, the said bundle tion of that bundle information and will formu-
could then be executed. Bundle will register the late some specifications such as Import-Packet,
service it provides in the OSGi platform and bundle Export-Packet, Bundle-Activator, Import-Service,
could also make a request upon OSGi platform Export-Service, etc.
for the services provided by other bundles. Bundle life cycle are mainly managed by
Framework is an integral information service the OSGi framework, its life cycle could be di-
platform; its main function is to provide bundle vided into six states: INSTALLED, RESOLVED,
with an environment of execution and dynamically STARTING, ACTIVE, STOPPING and UNIN-
to adjust the bundle life cycle. When a bundle has STALLED. The flow chart of the bundle life cycle
been installed, it will register the service it provides is shown in Figure 5.
at the framework; however, when a bundle has After an OSGi bundle has been installed on
been suspended, the framework will dynamically the framework, the services it provides could
remove the services registered by the bundle. The be presented by a service. A service must be
structure of framework could be divided into sev- a well-defined interface service. When each
eral layering; they are respectively security layer, bundle provides a service, the framework will
165
Figure 5. Bundle life cycle

framework itself; comparing to standard services,
custom services belong to the part that are defined
and developed by the vendors on their own. Each
vendor could develop its own bundle and the
corresponding service to segment the market of
the product.
After the above-mentioned introduction, we
could know that OSGi possesses the component
administration function and allows the user to
proceed with administration on components from
remote site; each OSGi framework will even
monitor the privilege of each bundle and differ-
entiate the execution of each bundle. Based on the
viewpoints mentioned above, we may realize that
reserve a corresponding service reference; and,
OSGi provides a safe environment for software
when another bundle needs the service provided
execution. If each bundle has enough privilege,
by this bundle, an interrogation could be made
and the bundle could access the service provided
upon framework through LDAP (Lightweight
by the counter-party bundle among one another,
Directory Access Protocol) to acquire that service.
this will be helpful to the modularized program
Inside the framework, an effective operation and
design. OSGi itself provides automatic bundle
application shall be intercommunicated through
download, dynamic installation, registration and
a series of services.
bundle updating; in which, it is not necessary to
Concerning the services presently defined by
re-activate Java virtual machine and makes the
the OSGi specifications, those include: standard
operation easier and more user-friendly.
services and custom services, the two are shown
For more detailed OSGi, one could refer to
in Figure 6. Standard services are provided by
the specifications published by OSGi Alliance.
Figure 6. OSGi standard services and custom services
166
Release 4 has been developed to be the current TCP/IP, UDP, HTTP, XML and SOAP to construct
version, those who are interested in could visit the UPnP platform. UPnP features also include:
OSGi Alliance official web site to download.
• Media and device independence: UPnP
upnp (universal plug and play) technology could be applied on many
transmission media including Ethernet,
Universal Plug and Play (UPnP, 2009) is a protocol IrDA, RF (Wi-Fi or Bluetooth), IEEE 1394
suite formulated by UPnP™ Forum founded in and Power lines.
October, 1990. The protocol vision is to establish • User interface control: UPnP technology
a seamless connection between the home network has enabled the device vendors to control
(data sharing, communication and entertainment) the device and to inter-communicate among
and the all kinds of devices within the enterprise one another through the web browser.
network, and to simplify the practice of relevant • Operating system and programming
network. In brief, it is expected that as long as each language independence: Any operating
device (e.g., computer, TV, fridge, air-conditioner, system and program language could be ad-
alarm clock or lamps, etc.) has been connected to opted to construct UPnP products.
the network, all the already on-net devices could • Program control: UPnP architecture also
sense that there is a new device added in, these enables the conventional application pro-
devices could then communicate one another gram control.
and could be directly operated and controlled; no • Extensibility: Each UPnP product can
configuration is required for any device; one could have device-specific services be layered on
fully enjoy the advantage of Plug and Play. top of the basic architecture.
For example, assuming one has bought a printer
at home, if you expects this printer be shared and The basic components that the UPnP consists
be used by all in-house computers, you will have of are Service, Device and Control Point. Service
to install this printer and set it to be sharable by is the minimum control unit in UPnP; service pro-
all. After that, you will also have to install this vides operative actions and a set of status variables
network-shared printer on other computers; will to record the current status of this service; device
it cause too much trouble? However, if we adopt represents an UPnP device, the device is the facil-
UPnP as the intercommunication bridge for all ity that includes services. For example, a DVD
in-house devices, as long as both the home printer player provides video-broadcasting service; the
and computer support UPnP, the user only needs control point is able to control the device found
to hook up the newly-bought printer, then, all in the UPnP network.
in-house computers will notice that there is one In the following, we will give a simple in-
printer available; no any configuration is required troduction to UPnP protocol stack as shown in
and the printer becomes usable. This is exactly Figure 7:
the vision that UPnP wishes to realize. Making
the configuration of the home network environ- • HTTPU/HTTPMU: These two communi-
ment to be simplified could further enhance our cation protocols are extensions of HTTP.
living quality. They use UDP/IP to transmit the data, and
UPnP is a technology realized on an open they are used by SSDP as well.
point-to-point IP network of Plug and Play. It is a • SSDP: Its full name is Simple Service
type of web-based communication protocol; and, Discovery Protocol; it is a communica-
it adopts the present widely used standards like tion protocol built in HTTPU/HTTPMU; it
167
Figure 7. UPnP protocol stack
defines mainly how to make the services The way to get an IP address could be done
be found out; it includes how the control by using DHCP or using auto IP to get the
point could find out what services exist on network address.
the network and access to the relevant in- • Discovery: It allows the control point to
formation about these services; that the de- find out the service of its interest. When
vice itself claim what services it could pro- a device is added into the network, that
vide is also completed through this SSDP the device will broadcast its services to
protocol. the control point in the network; when a
• SOAP: Its full name is Simple Object control point adds into the network, it will
Access Protocol. It mainly defines how to search the device of its interest in the UPnP
use XML and HTTP to execute the remote network. Using SSDP via HTTP/HTTPMU
procedure call. completes all these actions.
• GENA: The full name is Generic Event • Description: When a control point has
Notification Architecture; it is mainly ad- found out the device from the step of
opted to process the transmitting and re- Discovery, because the control point has
ceiving of subscription and notification pretty limited knowledge about that de-
messages. vice, if it intends to further understand the
As to the communication between the device device function or to interact with that de-
and control point, it could be separated as the vice, it will use the device XML address
following six phases: obtained in the Discovery step to acquire
the XML document of the device descrip-
• Addressing: It is the initialization step for tion. This document includes the device
a device to add into the UPnP network. In name, serial number and manufacturer, the
order to intercommunicate with the com- actions and status variables provided by
ponents in UPnP network, each control the device and the URL web address for
point and device must have an IP address. controlling this device.
168
• Control: When a control point intends to video films via the network. Therefore, as long
control a device, it will transmit a proper as the video facilities are conformable to DLNA,
control message to the service URL ac- they could be directly connected, synchronized
cording to the information acquired from and adopted for data transmission to one another
Description phase, this control message is without any driver and any interconnecting device.
expressed in XML format by using SOAP Till the year of 2008, DLNA alliance members
protocol. When a device receives this mes- have been reaching 245 (vendors), it includes
sage, it will act to change the correspond- quite a few manufacturing leaders of electronic
ing variables and then forward back to the products such as HP, Intel, Microsoft, IBM, Pa-
control point; if this action fails, it will re- nasonic, LG, Philips, SONY, Toshiba, Motorola,
turn an error code. Nokia, Samsung, etc. Till September of 2008,
• Event: A control point could subscribe the there are over 3000 kinds of products that have
status variable of its interesting device. passed DLNA certification.
When the status on a device changes, the The product specifications defined in “In-
device will issue an Event message, and teroperability Guidelines, Version 1.0” by DLNA
use GENA protocol to pass this message in June, 2004 contain the key components as
back to the control point. This Event mes- follows:
sage is also expressed in XML format.
• Presentation: If that device has a web GUI • DMS (Digital Media Server): It provides
for web page presentation, then the control the functions of media file access, record-
point will load that page into the browser, ing/producing, storing and being used as
and then the control point could inspect the a source, it is just like a multimedia file
device status and control the device from server. This type of device includes set-top
that web page. box, video (DVD) player, PC with built-in
media server function, broadcast receiver
The above is a simple introduction of UPnP; it and HD-embedded home theatre.
describes the operating environment and features • DMP (Digital Media Player): This device
of UPnP, UPnP protocol stack and every single generally represents a device that could
phase when being used. search and play or output any media file
provided by DMS; these types of devices
dlna (digital living network alliance) include TV, printer, home theatre, multi-
media cellular, PDA and some specified
DLNA is the abbreviation of (Digital Living game terminal machine. “Interoperability
Network Alliance, 2009). It is an alliance or- Guidelines, Version 1.5” is the recent ver-
ganization founded in 2003, which consists of sion published in March of 2006, and, it
the vendors from consumer electronic products, has been partially expanded in October of
mobile phones and computers. The organization the same year. Inside the v1.5 specification,
goal is to establish a set of standard specifica- several product specifications have been
tions that could allow the video/audio facilities added in addition to the previous DMS and
of different manufacturers and different types to DMP. It includes:
be inter-connectable and mutually adaptive to ◦ M-DMS (Mobile Digital Media
one another, and to realize the digital life for the Server): It is the server defined spe-
customers. As long as the user stays anywhere cifically against the electronic prod-
at home, he could access the photos, music and ucts like handheld mobile phone that
169
Figure 8. DLNA certified logo (Source: DLNA

is smaller in size, light in weight and
Website)
portable; the multimedia format it
supports is somewhat different from
that of the general DMS.
◦ M-DMP (Mobile Digital Media
Player): It is also designed specifi-
cally for the player of mobile device;
the multimedia format it supports is
somewhat different from that of the support and optional support. The graphic file for-
general DMP. mat of required support is JPEG, audio format is
◦ DMC (Digital Media Controller): LPCM, video format is MPEG-2; and, the graphic
When used as a remote device, it file formats of optional support are PNG, GIF and
could search the multimedia files that TIFF, audio compression formats are AAC, AC-3,
could be played on DMS and could MP3 and WMA9, the video compression formats
assign the DMP either to play that contain MPEG-1, MPEG-4, AVC, WMV9, etc.
multimedia file or to control the mul- For product test and certification, DLNA
timedia file uploading or download- provides standard specification and the software
ing to DMS. In addition to remote program that allows each vendor to do the self-
device, an intelligent device equipped test. The vendor could do the self-test first and
with basic operation interface could verify its normal execution, then, sent to DLNA
also be used as DMC. certifying organization for further certification.
◦ DMPr (Digital Media Printer): A After passing the certification, vendor will be
DMPr printer could provide printing awarded DLNA certification emblem as shown
function under the DLNA network in Figure 8.
framework. Basically, the DMPr Every single famous consumer electronic
printer function is similar to the tradi- product manufacturer has seen the DLNA vi-
tional USB printer. sion relatively promising. The currently adopted
method for solving the mutual communication
Concerning the mutual detection among one among the facilities is also suitable, however, the
another facilities, the UPnP standard introduced present DLNA standard specification may be still
in previous section is adopted. DMP facility could in some way not enough, such as the Digital Right
search the mutually matched DMS facilities on Management (DRM) of the multimedia file has
the network via UPnP mechanism, after successful been in great short, and at the same time, it also
connection, one could proceed with the following lacks security protection mechanism; it could
playing or data transmission actions. These actions be easily attacked by hackers or virus and thus
are all automatic, the user needs not to do extra causes the hidden concerning about the whole
configuration. As to the media content discovery, DLNA environment. Since DLNA has formulated
it adopts also the UPnP mechanism. Presently, the key product categories and network protocol
it adopts HTTP standard protocol regarding the framework, this allows the products made by
transmission. RTP will be then added in the future different vendors to have a mutually communi-
version and be used as the protocol for performing cative interface and platform, and solidly makes
audio and video streaming. the vision of digital home/vehicular network be
In the multimedia formats that DLNA supports realizable in the coming years of near future.
for transmission, it is divided again into required
170
inTerworking archiTecTure device outside the house to transmit SIP message

and design to remotely control the in-car devices.
The applied network framework is shown in
design of osgi gateway Figure 11.9. In order to let the user be able to
that supports sip communicate with each other while using devices
of different communication protocols (UPnP or
In a car, there could exist many kinds of network SIP device), except the OSGi Service Bundle that
devices and Networked Appliances (NAs), the OSGi already supports, one must depend on SIP
communication protocols adopted by different Service Bundle and Bridging Bundle (Bushmitch
devices are also not all the same; in order to et al., 2004; Brown et al., 2006) that have been
make in-car NAs to be able to inter-communicate provided additionally by OSGi platform. OSGi
with each other and to have a unified interface to Service Bundle provides the methods of device
administer and control NAs, OSGi Alliance has discovery and communication; the major func-
formulated a open service platform named OSGi tion of SIP Service Bundle is to enable the device
Service Platform (abbreviated as OSGi SP) to as- and the bundles to communicate with each other
sist in administering and controlling in-car NAs through SIP and to fulfill the functions of device
(such as X.10 device, HAVi device, Jini device, mobility and service mobility, etc. Besides, SIP
UPnP device, etc.). Service Bundle could be used to bridge two or
After having OSGi SP, in order to allow the more than two OSGi RGWs on the cars to make the
user to be able to remotely control the in-car NAs, RGWs to be communicative with each other.
one must make some updating and modifications After having SIP Service Bundle and Bridging
on the OSGi SP; to allow the user to connect via Bundle, an outside user could use his SIP device
the network to OSGi Residential Gateway (RGW) to transmit SIP message to remotely control the
even out of car and to further control the in-car NAs on the car. Firstly, after the outside user
NAs (the OSGi SP is practically implemented sends out the SIP message, the message will be
on the OSGi RGW). Here, we will discuss the sent to OSGi RGW via the SIP Proxy Server on
technology regarding the user’s utilization of SIP the Internet, and then gone through the SIP proxy
Figure 9. OSGi residential gateway supporting SIP
171
Figure 10. SIP service bundle(Brown et al., 2006)
server provided by SIP Service Bundle, the SIP a SIP service. After the SIP bundle registering,
message will be sent to a SIP UA. The SIP UA other bundles could then use this SIP service to
could probably be the SIP UA built in SIP device, register to become a SIP device. Then, the outside
it could also be the virtual SIP UA offered by SIP user could control the in-car SIP device via the
Service Bundle. The former will directly send the SIP message.
SIP message to that SIP device and let the SIP SIP Service consists mainly of three objects;
device process the received messages by itself; they are respectively SIPServer, SIPDevice and
the latter will be forwarded to Bridging Bundle, SIPUserAgent.
the SIP message will be translated into certain SIPServer – SIPServer, this object represents
specific message that is able to control the NAs the SIP Service server (proxy/registrar) in OSGi
(e.g., UPnP message format). RGW. If being served as a proxy server, it will ac-
In the following section, we will sequentially cept the registration message sent from SIP device
introduce the functional components and their or bundles, and send the received SIP message
operation modes of SIP Service Bundle and to the registered SIP UA. In addition, SIPServer
Bridging Bundle. could serve as a registrar server, allowing SIP
device or the non-SIP device of SIPUserAgent
SIP Service Bundle to register with the registrar, and providing some
additional ways for the non-SIP device to be able
As shown in Figure 10, SIP Service Bundle exists to register with the registrar server located at the
in a form of OSGi bundle within the OSGi frame- external network. Using SIPServer object could
work; that responsible for providing SIP support let SIP Service Bundle have the following several
for OSGi Bundles and the device that registers functions:
in OSGi is its major function. The features of
SIP Service Bundle contain service registration, 1. Asking a SIPUserAgent to serve as a virtual
messaging, event subscription/notification and SIP UA: When there is a non-SIP device
the complete SIP proxy and SIP server func- that intends to register with a registrar
tions. Once the bundle activates, it will make a server, the SIP Service Bundle will request
registration in OSGi service registry to become a SIPUserAgent to serve as a virtual SIP
172
UA. A virtual SIP UA has all SIP capabili- virtual SIP device. This virtual SIP device is just
ties, but it is not a physical SIP device. Once a bundle that requests a SIPUserAgent function.
SIPUserAgent has been produced, it could When a device has finished the registration, the
automatically register with the registrar SIP Service Bundle will produce a SIPDevice
server. object for that device to represent it. This will
2. To acquire a SIPDevice object that repre- allow other bundles in the OSGi framework to be
sents that device for any virtual or physi- able to control the device which the SIPDevice
cal SIP device that has registered with the represents.
OSGi framework: the SIP server must be SIPUserAgent: When a non-SIP device
responsible for doing a service registry on intends to register with a registrar server or
the OSGi framework for those SIP devices communicate with other SIP devices, it must
that have registered with the SIP server, to support the function of SIP UA. The SIP device
register them as the SIPDevice service. After itself supports SIP UA, so it could register with
that, the other bundles in OSGi framework registrar server or communicate with other SIP
could then use those SIP devices. devices directly. However, a non-SIP device (such
3. All of the SIP Service Bundle relevant events as air-conditioner, fridge or lamp) must ask SIP
that happened in OSGi framework will notify Service Bundle for a SIPUserAgent to serve as a
with the SIP Service Bundle. For example: virtual SIP UA in order to register with registrar
the devices registering/unregistering. to become a virtual SIP device or to communicate
with other SIP devices.
SIPDevice: SIPDevice, this object could
represent a physical SIP device and could also Bridging Bundle
represent a virtual SIP device. A SIP device could
register with the registrar to become a physical In order to perform the message translation
SIP device; a non-SIP device could register as a between two devices that adopt different com-
Figure 11.SIP-UPnP bridging scenario (Brown et al., 2006)
173
munication protocols, bridging is a compulsory Please refer to Figure 11 for the flow chart of this
process. Bridging is a mechanism of application- example. This type of bridging is applied on the
layer proxying. For example, translating the SIP same OSGi platform; and for the other type of
message into UPnP message or translating the application, it belongs to two devices that adopts
UPnP message into the SIP message. The software different communication protocols; about how to
developer could use the Application Programming mutually communicate with each other across the
Interface (API) provided by OSGi platform to different OSGi platforms. In the next section, we
import the required package (such as org.osgi. will briefly introduce how the bridging operates
service.upnp or org.osgi.framework), he could across the OSGi platforms.
then develop his own Bridging Bundles (such as
SIP-UPnP Bridging Bundle). inter-gateway bridging
Bridging Bundle is an OSGi bundle. It pos-
sesses the functions of asking a SIP UA from In the previous section, we have introduced how to
the SIP Service Bundle and could also ask the enable the mutual communication between the SIP
OSGi Service Bundle for an UPnP function. So, and UPnP device via an OSGi RGW. Therefore,
the devices that adopt different communication in this section, what we will discuss is another
protocols could intercommunicate with each other situation of application. Assuming the user owns
via different types of Bridging Bundles. some UPnP devices and an OSGi RGW in his
One of the bridging examples is to utilize a por- house. When he drives his car which equipped
table SIP mobile light control device to remotely with an OSGi RGW to travel, if he wishes to ac-
observe and control the in-car UPnP desk lamp cess the data in his home UPnP device, he must
status. Here, we assumed the desk lamps are all rely on the assistance of the SIP Service Bundle
UPnP devices. This kind of application example and Bridging Bundle embedded on the present
is to elucidate that no matter where the users are, mobile RGW on his car and the in-house home
they could remotely control the in-car devices via RGW in order to translate the SIP and UPnP mes-
the network. This SIP lighting controller could sages properly.
acquire the UPnP desk lamp relevant informa- As shown in Figure 12, if a user owns an UPnP
tion from the vehicular UPnP network environ- DVD player at home, and this DVD video record-
ment via Bridging Bundle. The controller cannot ing/player has registered with the SIP registrar
directly use the infrared technology to control server in the SIP Service Bundle to become a
the desk lamp; on the contrary, it has to depend virtual SIP device. Assuming the user wishes to
on the Bridging Bundle to interrogate the UPnP use the UPnP TV on his car to connect back to
Service. The controller could also subscribe the the DVD video player at his home to watch the
status changes with respect to every desk lamp. film stored in the DVD player.
Therefore, Bridging Bundle will listen the event First, the user will operate this UPnP TV, the
messages of all in-car desk lamps and utilize the TV sends out UPnP message (SOAP) to the UPnP
SIP NOTIFY message to forward the received Service Bundle. This UPnP Service Bundle is
event message to controller. In order to control responsible for providing the relevant informa-
this UPnP light, controller will send out SIP tion of the service and device that exist on every
MESSAGE to Bridging Bundle, and the Bridging registered UPnP device. UPnP service will pass
Bundle will translate the SIP MESSAGE contents the message to OSGi service registry via service
into the function call of UPnP Service API; in API. OSGi service registry will then pass the UPnP
the end, it will produce SOAP control message control message to SIP-UPnP Bridging Bundle.
and send them to UPnP light (the desk lamp). Therefore, UPnP message will be translated into
174
Figure 12. Inter-Gateway bridging example
SIP message in SIP-UPnP Bridging Bundle, and, sip and upnp mapping
be sent to virtual SIP UA that the UPnP TV has
registered on the SIP Service Bundle. In the end, UPnP has adopted many standard communication
via the SIP Service Bundle on the mobile RGW, protocols in order to make it compatible with the
the SIP message is sent back to SIP Service Bundle present network communication protocols; there-
on the home RGW. The processing sequence of fore, UPnP could be a cross-platform protocol.
this message in the home RGW is contrary to that Except the often heard protocols – IP, TCP, UDP
in the mobile RGW, and the SIP message will be and HTTP, in the following, some examples will
translated into the UPnP message in XML format be given regarding the not-often-heard UPnP
inside the SIP-UPnP Bridging Bundle of the home protocols:
RGW, and sent to UPnP DVD player. Therefore,
the user could establish the RTP package route SSDP (Simple Service
between mobile RGW and home RGW, and pass Discovery Protocol)
the RTP stream from the UPnP DVD video player
located at the home RGW to the UPnP TV located It is used when the UPnP intends to execute the
at mobile RGW. This is an example of integrating Discovery phase. It is exactly the timing when
home network and vehicular network through the control point of the home/vehicular network
inter-gateway bridging. (exactly the above-mentioned RGW) interrogates
175
the device of its interest about what services use his SIP device to send out SIP SUBSCRIBE
are available or when a device is added into the message to RGW, and the RGW will use the UPnP
UPnP network, it will automatically broadcast device address acquired during the Discovery
the services it provides to inform the control phase to execute the Description phase by using
point or RGW. HTTP (please refer to UPnP introduction). That
UPnP device will utilize HTTP to return the de-
SOAP (Simple Object Access Protocol) tailed information in XML format to OSGi RGW,
and the Bridging Bundle in OSGi RGW will then
It is used when the UPnP intends to execute the translate this message into SIP NOTIFY message
Control phase. Once the control point or RGW has and send it back to the SIP device.
acquired the relevant description of certain specific In addition, a user could subscribe the device
device, it will be known that what actions could status from OSGi RGW in advance. When the
be applied on that device, so, the control point or status of that device changes, the OSGi RGW has
RGW will use the SOAP to issue commands to to inform the SIP device. First, the user will send
control the status of that device. out SIP SUBSCRIBE message to OSGi RGW via
his SIP device, once the status of the UPnP device
GENA (General Event that we intend to subscribe has changed, the UPnP
Notification Architecture) device will use GENA protocol in XML format to
send the event message back to OSGi RGW; then,
It is used when the UPnP intends to execute the the Bridging Bundle in OSGi RGW will translate
Event phase. The user could subscribe the device this message into SIP NOTIFY message and send
it intends to control via the control point or RGW; back to the SIP device to inform the status change
if the status of that device has changed, the device of that UPnP device.
could use GENA to issue the event notification
to the control point or RGW to inform its present MESSAGE
status of that device.
The mapping way between SIP and UPnP is to After the SIP device has acquired the detailed
utilize SIP message to control UPnP device and the description of the UPnP device one wishes to
event notification. Therefore, regarding the mes- control, one could send out the SIP MESSAGE
sage interchange method between SIP and UPnP, to OSGi RGW. The Bridging Bundle in OSGi
except the above-mentioned UPnP communica- RGW will again translate this SIP message in
tion method, it also includes three types of major XML format and use SOAP protocol to control
SIP communication ways; they are respectively UPnP device.
SUBSCRIBE, NOTIFY and MESSAGE. In the
following, we will introduce how SIP and UPnP integrating ims and dlna
could achieve the messages translation. connected home
SUBSCRIBE/NOTIFY The concept of Connected Home is originated from

the common vision of DLNA alliance and UPnP
When an outside user wishes to use the SIP device organization. No matter what brand the home
registered in the RGW to connect back to the OSGi devices belong to, they could communicate with
RGW on the car and ask to access some functions each other, share the multimedia data, the home
provided by the UPnP device of interest and the surveillance and remote access service. However,
detailed information. At this moment, the user will in order to let a user access remotely the device
176
at home/car and enjoy the guarantee of security higa (home ims gateway)
and quality of service at the same time, it must
need a reliable system to achieve the remote ac- HIGA is currently formulated by the two organiza-
cess service. tions of (Home Gateway Initiative, 2007; ETSI
In order to achieve this goal, Ericsson has TISPAN, 2008). It is a standard of NGN home
proposed a framework (Fasbender et al., 2008) gateway; it can be regarded as a gateway on the
that has combined the advantages of IMS, DLNA car as well. HIGA is a functional component op-
and UPnP as shown in Figure 13. The user does erating on RGW. It is responsible for collecting
not need to purchase extra specific device; this the relevant information of the in-house/in-car
architecture utilizes the existing IMS framework users, devices and services, and managing the IMS
to provide the authentication and authorization of session for the home/ in-car devices that do not
the user identity, message route, and to establish support IMS function. With the HIGA, RGW then
the secured multimedia sessions and QoS guar- contains an ISIM card (IMS Subscriber Identity
antee. Besides, the design of this framework is Module Card), it allows the in-car devices to ac-
fully conformable to the present standard of the cess the services provided by IMS network via
consumer electronic products for providing mul- the RGW. HIGA will also perform the transla-
timedia data sharing and service delivering. The tion between the protocols adopted by the in-car
core component in this framework is the Home devices and IMS/SIP.
IMS Gateway (HIGA) functional component HIGA could use IMS SIM card to register with
that Ericsson has continuingly developed on the IMS core network in the secured authentication
Residential Gateway (RGW) since 2005. method. The in-car devices could communicate
Figure 13. High-level remote access architecture (Fasbender et al, 2008)
177
with the IMS network through the HIGA which is and inter-changed contents between RAC and
equipped with the functionalities of Back-to-Back Vehicular Network.
UA (B2BUA) and SIP UA. For example, there Regarding all of the devices included in this
is a SIP VoIP phone on the car, this phone could framework (Remote UE, Residential Gateway
then register with HIGA by using the SIP UA it with HIGA and DLNA Devices), one could refer
owns by itself, the B2BUA in HIGA could then to Figure 14 for the block diagram of functional
translate this SIP message into the IMS specified components that are supposed to be equipped.
message and forward to IMS network. HIGA could be implemented on any in-car device,
In reality, this framework is an extension of but practically speaking, implementing directly
UPnP Remote Access standard; making it be the HIGA on RGW will be the most simple and
able to support IMS Remote Access and allow- of highest feasibility because RGW itself supports
ing the concept of DLNA Connected Home to Network Address Translation (NAT) and Firewall
be extended to the outside of the home. The two (FW) functions. If the HIGA function is integrated
major functional modules of UPnP Remote Ac- into RGW, when a user at outside wishes to ac-
cess framework are Remote Access Transport cess the in-car devices, this HIGA will be like
Agent (RATA) and Remote Access Discovery an end point of the signaling delivering in the
Agent (RADA). These two kinds of modules will IMS network; and, in the vehicular network, the
be implemented on the Remote Access Client HIGA will be like an UPnP device that intends to
(RAC) and the Remote Access Server (RAS). send the message to the in-car device. In the very
RATA is used to establish a secured communi- timing that makes the vehicular network and next
cation channel between the Remote UE (User generation all-IP network to be closely integrated,
Equipment) and the Vehicular Network; RADA not only it could provide diverse IMS multimedia
is used to synchronize the UPnP device message integrated services to the user, it could also al-
low the user to access any device of his own no
Figure 14. Functional architecture
178
matter anywhere, this will make our lives more Phase 1: Connection Request
convenient and digitized.
Assuming Alice is traveling outside the country,
ims approach for remote access she wishes to access the video film stored in the
NAS via her mobile device (in this example, it is
With the supports of DLNA and UPnP in the the Remote UE); the Remote-Access application
vehicular communication environment, we could on her device will send out the IMS INVITE mes-
establish the remote access session between the sage to the HIGA on her car. When the INVITE
Remote UE and the RGW that supports HIGA via message passes IMS network, the user’s IMS
the IMS cooperation. Assuming the Remote UE home network will add a P-Asserted-Identity into
has a digital multimedia renderer (DMR) which the INVITE message. At this moment, the HIGA
supports DLNA, the Network Attached Storage will cross-compare the P-Asserted-Identity and
the permitted user identities to verify the Remote
(NAS) supports the function of digital multimedia
UE. The SDP (Session Description Protocol)
server (DMS), and both the Remote UE and HIGA
packed in IMS message will be used to inform the
have registered with IMS and could connect to
RAS located in the HIGA and the RAC inside the
IMS network. Therefore, the outside IMS user
Remote UE of the IP address and the port number
could connect to the RGW on the car via the IP
of the remote access tunnel. The SDP parameters
Multimedia Public Identity (IMPU) of the HIGA. are also used to negotiate the administration of
The flow example that uses IMS to proceed with the virtual private network tunnel (VPN tunnel)
remote access could be divided into six phases as and encrypted keys that are established between
shown in Figure 15. Remote UE and HIGA.
Figure 15.Session setup flow between remote UE, HIGA and NAS (Fasbender et al., 2008)
179
Phase 2: Peer-to-Peer VPN Setup multimedia contents are available, and could
over the IMS Media Plane choose to playback the mentioned multimedia file
by downloading or via streaming method. If one
Once the Alice’s device has completed its authen- wishes to forward any multimedia data through
tication and authorization procedures, the device the VPN connection built in phase 2 in the IMS
has then established a secured media transmission media plane, one could forward the data by using
mechanism with HIGA. The connection method HTTP (the DLNA default transmission protocol)
adopted by Ericsson will set up the connection or RTP (DLNA extra supported transmission
flow according to the IMS standard and will protocol).
make HIGA as the VPN server. Therefore, after
the tunnel has been established, the connection Phase 5: IMS Media
between the Alice’s device and the in-car NAS Plane QoS Upgrade
will be exactly like the UPnP connection at the
local terminal. UPnP does not support QoS management in
LAN. But, after integrating the telematics and IP
Phase 3: UPnP Discovery Multimedia Sub-system, one could use the QoS
supported by IMS standard to make it be able to
The original design of UPnP is to allow the device support the QoS control and administration be-
to be able to mutually communicate with each other tween the remote UE (Alice’s device) and HIGA.
in a LAN, if one wishes to extend the UPnP to a If it is expected to update some QoS parameters
WAN, one will surely be confronted with some for Alice’s device, the device could send out re-
problems. For example, UPnP is using a message- INVITE or UPDATE message to the IMS network
interchanging method of multicast to find out the and HIGA to update the QoS status.
UPnP device. However, these multicast packets
will be discarded by the routers located on the Phase 6: Content Playout
Internet. Nevertheless, the remote access method
formulated in UPnP Remote Access standard could In the last, the multimedia data chosen by Alice
allow the UPnP RAS filtering device to dig out could then be played on her mobile device.
some relevant messages, and using unicast method With the IMS-based remote access support,
to transmit the message of device discovery to the it could make up some defects that the original
Alice’s device located outside. UPnP remote access lacks. Those defects include
If adopting UPnP RADA mechanism, it could some security holes, the authentication of remote
make the RAC of the Alice’s mobile device to user and the QoS provision. Therefore, the DLNA/
be synchronized with the RAS on HIGA, then, UPnP applications can be extended to the main
Alice could use her device to access those avail- system of NGN, it could make the vehicular
able multimedia servers and the UPnP services network and mobile communication network to
provided by the server. RADA could be used to be more diverse and allow the user to have richer
dynamically inform the Alice device of what new experience in the world of multimedia network.
devices have been added on the car.
Phase 4: Content Selection conclusion
Alice chooses the NAS, using UPnP Content In vehicular network applications, in addition to
Directory Service (CDS) to browse what the solving the mutual communication issue among
180
the devices in every single vehicular network, reFerences

one needs to consider the connectivity of several
vehicular networks. This will allow us to make a Alliance, O. M. (OMA). (2009). Retrieved from
connection back to the RGW on the car no matter http://www.openmobilealliance.org/
where we are to achieve the goals of remote access Brown, A., Kolberg, M., Bushmitch, D., Lomako,
control and multimedia data sharing, etc. But, the G., & Ma, M. (2006). A SIP-based OSGi Device
above applications do not possess enough security Communication Service for Mobile Personal
authentication mechanism with respect to the ve- Area Networks. In Consumer Communications
hicular network; besides, it lacks certain degree of and Networking Conference(CCNC).
QoS guarantee regarding the multimedia contents
that the user wants to transmit. So, in recent years, Bushmitch, D. Lin, Wanrong., Bieszczad, A., Ka-
there are some organizations of standardization plan, A., Papageorgiou, V., & Pakstas, A. (2004).
that are formulating the integration framework A SIP-based device communication service for
between the digital home and IMS. The design of OSGi framework. IEEE Consumer Communica-
this framework requires totally no modification tions and Networking Conference2006.
on the present existing terminal equipments (such
DLNA. (2009). Retrieved from http://www.dlna.
as IP devices, SIP devices, Bluetooth devices
org/home
and UPnP devices), it only requires extra adding
of HIGA functional component on RGW, it will ETSI TISPAN. (2008). Retrieved from http://
enable the in-house/in-car devices to access IMS www.etsi.org/tispan/
services via RGW, and it also allows the user to
Fasbender, A., Gerdes, M., Hjelm, J., Kvarnström,
connect to the RGW anywhere via the reliable
B., Petersson, J., & Skog, R. (2008). Virtually
authentication mechanism provided by IMS to
at home: High-performance access to personal
perform a further control on the in-house/in-car
media. Ericsson Review, 2.
devices.
Because the most time of our lives are spent 3GPP. (2008). IP Multimedia Subsystem (IMS);
either out of home, in the office or on the way Stage 2, TS23.228 v8.6.0.
to somewhere, so, except to integrate the home/
3GPP. (2008). IP Multimedia (IM) Session
vehicular network with next generation all-IP
Handling; IM Call Model; Stage 2. TS 23.218
network framework, on the other hand, DLNA
v8.3.0.
standard is expected to integrate the home network
and telematics of the vehicle industry. We could Home Gateway Initiative. (2007). Retrieved from
treat a car or any public transportation as a mobile http://www.homegatewayinitiative.org/
home network, and could also regard a car as a
Open Service Gateway Initiative (OSGi). (2009).
device for remote accessing to home network;
Retrieved from http://www.osgi.org
allowing the user to connect back home to access
any multimedia data while he is on a vehicle and Rosen, J., Schulzrinne, H., Camarillo, G., John-
to administer or control the devices in house. By ston, A., Peterson, J., Sparks, R., et al. (2002). SIP:
allowing the user to be able to interact with the Session Initiation Protocol. IETF RFC 3261.
home/vehicular network no matter at anywhere,
Third Generation Partnership Project [3GPP].
this could then connect more closely our network
(2009). Retrieved from http://www.3gpp.org
world together.
UPnP Forum. (2009). Retrieved from http://www.
upnp.org
181
Section 5
Vehicular Ad Hoc Networks and
Delay Tolerant Vehicular
Networks
183
Chapter 12
MAC Protocols in Vehicular
Ad Hoc Networks
Chih-Yung Chang
Tamkang University, Taiwan, R.O.C.
absTracT
With the rapid development of wireless technologies, the Vehicular Ad Hoc Networks (VANETs) have
recently received much attention. VANETs technologies aim to ensure traffic safety for drivers, provide
comfort for passengers and reduce transportation time and fuel consumption with many potential ap-
plications. The achievement of these aims highly relies on efficient MAC protocols which determine
the performance of packet transmission in terms of success rate, delay, throughput and bandwidth
utilization. This chapter reviews the existing MAC protocols developed for VANETs. Initially, the IEEE
802.11p and DSRC standard are reviewed. Three TDMA-based MAC protocols, called CVIA, VeSOMAC
and D*S, are then introduced. In addition, three MAC protocols that cope with the emergency-message
broadcasting problem are proposed. Finally, a reliable MAC protocol which is developed based on the
cluster topology is reviewed.
inTroducTion and throughput issues on the demand of Internet ac-

cess in a multi-hop manner. The VeSOMAC MAC
This chapter reviews the existing MAC protocols protocol intends to reduce the transmission latency
developed for VANETs. Initially, the IEEE 802.11p for broadcasting a message over the VANETs. The
and DSRC standard are reviewed in Section 2. Three D*S introduces the data access MAC scheduling
TDMA-based MAC protocols, called CVIA, VeSO- protocol which aims to develop a scheduler in Road
MAC and D*S, are then introduced in Section 3. The Side Unit (RSU) so that those upload and down-
CVIA MAC protocol mainly concerns the fairness load requests can be satisfied as more as possible.
In Section 4, three MAC protocols proposed for
broadcasting the emergency message are depicted.
DOI: 10.4018/978-1-60566-840-6.ch012
MAC Protocols in Vehicular Ad Hoc Networks
Figure 1. Protocol stack relating to OSI model

These MAC protocols aim at avoiding the packet
collision and achieving high success rate. Finally,
Section 5 reviews a reliable MAC protocol which
is developed based on the cluster topology. The
conclusions are finally given in Section 6.
ieee 802.11p and dsrc

sTandards
DSRC (Dedicated Short Range Communications)

(ASTM International E2213-03, 2003) is a well
known standard supports both Public Safety and
Private operations in roadside to vehicle and
plication layer includes the fragmentation and
vehicle to vehicle communication environments.
defragmentation of data application service and
DSRC standard at 5.9 GHz band is projected to
service primitive for a variety of applications,
support low-latency wireless data communications
including electronic toll collection, emergency
between vehicles and from vehicles to roadside
warning system, vehicle safety service, commerce
units. The DSRC specification is meant to be an
transactions via cars, electronic parking payments,
extension of the IEEE 802.11 technology into the
probe data collection and so forth.
outdoor high-speed vehicle environment. In fact,
The standard of DSRC is comprised of IEEE
the Physical Layer (PHY) of DSRC is adapted
802.11p and IEEE 1609 family. Figure 1 shows the
from IEEE 802.11a PHY based on Orthogonal Fre-
correspondence between DSRC and OSI models.
quency Division Multiplex (OFDM) technology.
IEEE 802.11p (Draft 7.0, 2009) is a draft amend-
Moreover, the Multiple Access Control (MAC)
ment to the IEEE 802.11 standard to add Wireless
layer of DSRC is very similar to the IEEE 802.11
Access in the Vehicular Environment (WAVE).
MAC based on the Carrier Sense Multiple Access
As an extension of IEEE 802.11, IEEE 802.11p
with Collision Avoidance (CSMA/CA) protocol
defines how data exchange between high-speed
with some minor modifications.
vehicles and between the vehicles and RSU in
DSRC is meant to be a complement to cellu-
5.9 GHz (5.85-5.925 GHz) band. IEEE 1609 is a
lar communications by providing very high data
higher layer standard based on IEEE 802.11p. The
transfer rates in circumstances where minimizing
IEEE 1609 family is consisted of IEEE 1609.1
latency in the communication link and isolating
(2006), IEEE 1609.2 (2006), IEEE 1609.3 (2007)
relatively small communication zones are im-
as well as IEEE 1609.4 (2006).
portant. DSRC protocol layer is developed based
IEEE 1609.1 plays the role of WAVE Re-
on physical, data link and applications layers
source Manager which specifies a DSRC ap-
of traditional OSI model. In the Physical layer,
plication overlying WAVE and allows remote
DSRC defines physical parameters for uplink and
site applications to communicate with OBUs or
downlink communication and is mainly working in
RSUs. As a standard of application layer, IEEE
the 5.9 GHz band (U.S.) or 5.8 GHz band (Japan,
1609.1 conducts application-level information
Europe). In the data link layer, DSRC defines
interchanges.
frame format, frame wrapper and the procedures
IEEE 1609.2 supports WAVE security ser-
of MAC and Logical Link Control (LLC). Ap-
vices for applications. It defines secure message
184
Figure 2. The frequence map of DSRC
formats, specifies methods for securing WAVE Tdma-base mac

management messages and application messages proTocols on VaneTs
and processes secure messages of DSRC/WAVE
systems. In addition, it handles exception of A number of VANET MAC protocols have been
vehicle-originating safety messages and provides proposed based on TDMA system. Compared
administrative functions necessary for core secu- with probabilistic-based MAC protocol, an obvi-
rity functions. ous advantage of TDMA system is the ease for
IEEE 1609.3 provides WAVE devices and sys- avoiding collision and contention with smart as-
tems with WAVE networking services. It mainly signment for the transmission of each slot. This
defines WAVE networking services operated at subsection reviews some TDMA-based MAC
the network and transport layers. In other words, protocols developed for VANET.
IEEE 1609.3 represents roughly layers 3 and 4 of
the OSI model and the IP, UDP and TCP elements cVia mac protocol for
of the Internet model. Vehicular networks
IEEE 1609.4 supports frequency band coor-
dination and management within the MAC layer, An efficient TDMA-based MAC protocol can as-
supplements 802.11 MAC features as well as co- sign each transmission with a slot without collision
ordinates operations on both the Control Channel such that potential parallel transmissions can be
(CCH) and Service Channels (SCH). exploited. Korkmaz et al. (2006) proposed a new
Figure 2 depicts the Frequency band of DSRC. cross-layer protocol, called Controlled Vehicular
The frequency band used by DSRC ranges from Internet Access (CVIA), for vehicular Internet
5.855GHz to 5.925GHz and is divided into seven access along highways protocol. The proposed
10MHz communicating channels and one 5MHz CVIA protocol can be applied on the application
reserving channel. The 5.885~5.895GHz is control such as Internet access through RSU in a multi-
and safety emergency channel. The rest of band hop manner. In the network environment of CVIA,
is reserved for public safety or private operations. the service region is not fully covered by the
Furthermore, two specific channel pairs, channel deployed RSUs due to hardware cost constraint.
174 with channel 176 and channel 180 with chan- As a result, vehicles might need to exchange data
nel 182, are able to merge to be a 20MHz channel with RSU via multi-hop communication. Since
for special conditions. vehicles with longer distance to the RSU have
fewer opportunities to exchange their data with
the gateway, one major problem considered in this
research is the fairness. Other issues considered in
185
Figure 3. Concepts of slots and segments in CVIA
CVIA include collision avoidance and throughput segment closer to RSU. As shown in Figure 4, the
enhancement. data sent by TRiout
+1
was forwarded by forwarders
As shown in Figure 3, the proposed CVIA TRi , TRi and TRiin-1 subsequently.
in out
divides service region into a number of equal- Three major tasks should be performed by
sized segments. Segment Si closer to the RSU vehicles of some specific segment in the allocated
will be assigned with a smaller identifier number the time slot. Figure 5 gives an example for il-
i. The CVIA protocol assigns a specific time slot lustrating the basic concept of CVIA. As shown
to each segment so that vehicles belonging to the in Figure 5a, the TRiin ,next delivers the packet
segment can access the channel with collision. The train originating from segment Si+1 to TRiout .
objective of CVIA is to increase the end-to-end After, as shown in Figure 5b, local packets of the
throughput while achieving fairness in bandwidth
segments Si are gathered by TRiout in a collision
usage between road segments.
in avoidance manner. Finally, the TRiout of segment
In each segment, two vehicles TRi and Si creates a new packet train based on received
TRiout will be selected as temporary routers for packets and then sends it to TRiin-1 in Si-1 , as
collecting and then forwarding data packets. The shown Figure 5c.
TRiin takes charge of forwarding data sent by As shown in Figure 6, the CVIA algorithm
the vehicular in the neighboring segment with mainly consists of six phases, namely Inactive,
bigger ID whereas TRiout is mainly responsible Vehicle Position Update, Temporary Router Selec-
for collecting the packets to the forward of next tion, Intra-segment Packet Train Movement, Local
Figure 4. Packet movement via routers
186
Packet Gathering Phase as well as Inter-segment odd slots. All vehicles can be aware its segment
Packet Train phases. Figure 7 gives an example number according to its location information.
that would be used throughout this section for Therefore, each vehicle knows whether or not it
illustrating the detail operations of each phase. should be active for a given time slot. A vehicle
(a) Inactive Phase. Vehicles in this phase that stays in inactive phase will do nothing.
are not permitted to exchange data for avoiding (b) Vehicle Position Update Phase. This phase
collision and contention. For a given time slot, mainly exchanges the location information of
each vehicle can easily know whether it stays in all vehicles located in the active segment. The
inactive state or not. The reason is stated below location information will be further used in the
using the example shown in Figure 7. We assume later phases. Each vehicle should first checks if
that each RSU serves six segments Si, 1 ≤ I ≤ 6. it stays in active state for a given time slot. If it is
In addition, we assume that vehicles are able to the case, the vehicle should execute the operations
obtain their positions and synchronize their clocks defined in the Vehicle Position Update Phase. Let
through the equipped GPS. A set of time slots are time interval tu is the length of this phase. In this
divided into two sets, odd and even sets, depend- phase, each vehicle picks a Random Waiting Time
ing on the number labeled on the slot is odd or (RWT) from 0, tu - tPUP, where tPUP is the time dura-
even. The six segments are also divided two sets, tion required for sending a packet. After an RWT,
namely the odd set {S1, S3, S5} and the even set vehicles access the channel using the Distributed
{S2, S4, S6} Segments belonging to the same set Coordination Function (DCF) as defined in the
will be active at the same time slot. Herein, we IEEE 802.11 protocol.
assume that the even and odd time slots are allo- (c) Temporary Router Selection Phase. The
cated to even and odd segment sets, respectively. topology of the vehicular network is dynamic.
For example, vehicles in segments S1, S3, S5 are New temporary routers should be selected peri-
active and are permitted to exchange data in the odically, because the router may move to the next
Figure 5. Three important phases in CVIA
187
Figure 6. CVIA protocol state machine
segment. The selected routers are called TRiin ,next First, TRiin receives the packet train coming from
and TRiout ,next until they become active. The TRiin segment Si+1. Then it selects and announces the
is responsible for selecting new TRiin ,next and TRiin ,next and the TRiout ,next at the beginning of the
TRiout ,next after the position update phase. The next active slot. Final, TRiin forwards the packet
TRin uses two parameters, router lifetime and train to TRiout in the next active slot. As a result,
safe area, to choose the new routers. The router the router lifetime of TRiin should be carefully
lifetime represents the time which routers should calculated accordingly.
stay in the segment. Therefore, the router times The candidates of TRiout ,next and TRiin ,next
of TRin and TRout will be different because they should exclude the vehicles which will move
have different tasks as described below. out the segment during the router lifetime. The
The TRiout is responsible for three tasks in the farthest and closest candidates away from the
active time. First, TRout receives the packet train RSU will be selected as TRiin ,next and TRiout ,next ,
sent from TRiin . After, TRout collects the local respectively. Finally the TRiin will broadcast the
packets from all vehicles in the same segment. result of new inheritors.
Finally, TRout creates a new packet train and (d) Intra-segment Packet Train Movement
then sends this train to TRi-1 in which is located Phase. In this phase, a priority policy should be
in the next segment. Hence the router lifetime of designed to guarantee that the forwarding packets
TRiout should be long enough for executing the can be delivered to the Trout of the same segment.
three abovementioned tasks. The router lifetime The priority of TRiin transmission should be high-
of TRiin is quite different from that of TRiout . est since it is in charge of forwarding packets sent
188
Figure 7. Slot allocation of CVIA
from router of previous segments. To achieve this, access channel if it detects that the medium is idle.
TRiin only need to wait for a SIFS period as its In case that the medium is busy, the sender will
active slot starts. again wait for another random time period. In this
(e) Local Packet Gathering Phase. In this out
phase, TRi has the highest priority to transmit
phase, TRi-1 aims to receive the data packets of in this phase. To maintain the fairness among all
those vehicles in the same segment. All vehicles vehicles in different segments, a predefined size
in segment Si have the same priority for access- for local packets will be the constraint for data
ing channel. Therefore they access channel in a transmission in each segment. Since Trout is
contention-based manner. To avoid the occur- responsible for receiving all local packets in this
rence of collision, each vehicle should wait for phase, it can be aware whether or not the size
a random period of time. An awaken sender can of received packet is larger than the predefined
189
size. If it is the case, the Trout should broadcast emerging DSRC standards. A distinctive feature of
a message to terminate this phase. VeSOMAC (Yu et al., 2007) is its distributed design
(f) Inter-Segment Packet Train Movement with fast schedule reconfiguration for coping with
Phase. In the last phase, TRiout will create a new vehicular topology changes. The main contribution
packet by combining the packets collected in the of the paper is the design of a self configuring
Intra-segment Packet Train Movement Phase and TDMA protocol capable of inter-vehicle mes-
Local Packet Gathering Phase. The new packet sage delivery with short and deterministic delay
will be sent to the router of next segment. Then bounds. The most significant contribution of this
this phase will be terminated. All vehicles move protocol is that it allows a set operations running
back to the inactive state and wait for the next in a TDMA-based system without strictly clock
active slot. synchronization. The proposed Vehicular Self-
In summary, the CVIA protocol improves the Organizing MAC (VeSOMAC) assume that each
performance of transmission in the environment vehicle is aware of its own location and velocity.
which takes into consideration the vehicle-to-vehi- The following uses an example shown in Figure
cle and vehicle-to-roadside communications. The 8a to illustrate the basic concept of VeSOMAC.
packet collision ratio of CVIA decreases and the Assume that there are four vehicles, A, B, C, and
fairness among segments are achieved. However, D, move in order. After an emergency event (e.g.
the fairness is achieved between segments, rather an accident) occurred in front of the platoon A-B-
than between vehicles. A particular case that one C-D, the platoon head A periodically broadcasts
segment has light traffic demands and the other warning messages instructing other vehicles to
segment has heavy traffic demands, allocating slow down for avoiding collisions. Such warning
same bandwidth to all segments is not a good way. messages are to be forwarded by all vehicles across
In addition, the bandwidth utilization of CVIA can the entire platoon with minimum possible delivery
be further improved. To avoid packet collision, latency. With an example TDMA allocation with
each segment is assigned with an equal-length ac- arbitrarily slot placement as shown in Figure 8b,
tive slot. However, the constraints of packet size it will take three TDMA frames for delivering the
for different segment are different, which results messages to all vehicles in the platoon. However,
in unused bandwidth occurred in the segment farer with a possible VeSOMAC allocation, in which
to the RSU. As a result, bandwidth utilization can slots are allocated based on the vehicles’ relative
be improved. Besides, the slot-based approach locations as shown in Figure 8c, the delivery delay
requires accurate time synchronization which is can be significantly reduced. In this example, all
difficult to be implemented in the real world. messages can be delivered within a single frame.
This improvement can be much more pronounced
Vesomac mac protocol for larger platoons. This way VeSOMAC can ef-
fectively enhance highway safety by leveraging its
In a VANET, emergency message such as ac- ability to allocate slots based on location, speed,
cident information should be broadcasted within and other vehicular contexts.
an acceptable time period. To avoid the transmis- Information about allocated slots is exchanged
sion delay raised by collision and contention, a among the vehicles using a Bitmap Vector in each
TDMA-based system would be a better candidate packet header. Figure 9 depicts an example to il-
than probabilistic-based system. Fan Yu et al. pro- lustrate the concept of bit-map. The slot allocated
posed a novel Medium Access Control protocol to vehicle B is marked with blue ink while the
for inter-vehicular wireless networking using the slots occupied by all of B’s one-hop neighbors are
marked with green ink. Although these neighbors’
190
Figure 8. Basic idea of VeSOMAC
slots are shown with respect to B’s frame, each Using this header bitmap, a vehicle continu-
neighbor maintains its own asynchronous frame. ously informs its 1-hop neighbors about the slots
The middle of the bitmap vector represents B’s occupied by its 1-hop neighbors. By listening to
slot time. The bitmap vector is 4-bit long and each the bitmaps in all received packets, a vehicle can
bit represents the occupancy status of two slots detect the slot locations of its 1-hop and 2-hop
around B’s own slot. The major reason for using neighbors. This information can then be used for
one bit to represent two slots is that the neighbor’s choosing a slot which is non-overlapping with
slot can partially overlap with two contiguous the one and two hops vehicles’ slots.
slots of B’s frame in the asynchronous mode. For Slot allocation in VeSOMAC needs to satisfy
example, the ‘1’ in “+1” location indicates that at the following constraint:
least one of the two slots immediately following
B’s slot are already fully or partially occupied. 1. Timing constraint: This constraint indicates
Similarly, a ‘0’ in the “-1” location indicates that no two one-hop or two-hop neighbors’
that vehicle B perceives both the slots before its slots can overlap. Overlaps between one-
own slot to be free. The bitmap vector length is hop or two-hop neighbors cause direct and
a design parameter whose maximum value is the hidden collisions respectively.
frame slot count. In Figure 9, the frame size is 12, 2. Bitmap constraint: For 1-hop neighbors
whereas the bitmap length is 4, which can convey i and j, i’s chosen slot should be able to be
the occupancy information about only 8 slots. represented within the bitmap vector of j.
With a bitmap size 4, B is unable to represent the The same is applicable for vehicle j’s slot.
occupancy information about one of its neighbors’ In the asynchronous case, since each bit
slots-the one in extreme left. corresponds to two slots, this constraint
means that the slots of vehicles i and j can’t
191
Figure 9. In-band header bitmap for the asynchronous operation
be more than x slots apart, where x is the 4. The VeSOMAC protocol state machine
bitmap length. with all three constraints is presented in
3. Ordering constraint: If two vehicles i and Figure 10. The Stable state for a vehicle
j are geographical neighbors and i’s location indicates the time slot allocation matches
is ahead of j in the platoon, then i’s chosen the order of geographic location order of ve-
slot should be earlier than j’s slot in the time hicles. The Listen and Evaluate are transient
domain. The ordering constraint is optional, states. In the Listen state, vehicle overhears
and it is useful when the wireless messages the broadcasting message and look at the bit-
flowing from the front to the tail of a platoon map information. Then the vehicle chooses a
are more delay critical than the messages slot according to its physical location. After
flowing in the reverse direction. a vehicle chooses a slot through the Listen
state, it spends a preset (W) number of slots
Figure 10. State machine of VeSOMAC protocol
192
in the Evaluate state before getting into the can’t choose an empty slot, which might cause
Stable state. In the Evaluate state, the slot the new vehicle starvation. Therefore, how to
is evaluated for W frames to make sure that decide the frame size by estimating the density
the vehicle monitors its neighborhood activi- of the vehicle network will be an important issue
ties to decide if its own allocation became in the future works.
stable. If the assigned slot order is different
with the order of geographic locations of the data access mac scheduling
vehicle and its neighbors, it stays in Evaluate protocols for rsu
state and tries to change its selected slot.
When the state machines for all vehicles in Recently, vehicle-roadside data access has re-
a neighborhood reach the Stable state, the ceived considerable attention in VANET. VANET
protocol is said to have converged. is a special case in MANET (Mobile Ad-hoc
Network) and it is composed of vehicles with
The VeSOMAC proposes a distributed slot wireless communication devices and RSUs.
allocation mechanism to avoid the packet col- Vehicles can upload and/or download data from
lision between one-hop and two-hop vehicles, RSU. For example, drivers can get lots of traffic
and minimizes the delay of the event delivery or map information from RSU to improve the
by exchanging the Bitmap among neighboring safety in VANET. Therefore, RSU can be treated
vehicles. Furthermore, the proposed mechanism as a buffer point (or data island) which can store
can be applied in both time synchronous and time a variety of data including the value-added ad-
asynchronous scenarios. However, the efficiency vertisement, real-time traffic as well as digital
of the proposed mechanism is highly affected by map. As shown in Figure 11, the RSUs are often
the frame size, and thus the number of slots in deployed at the road intersections or areas with
a frame should be carefully determined. If the high traffic to improve the serving efficiency to
number of slots in a frame exceeds the vehicles, each vehicle.
the chosen slots may be non-continuous and the In VANET, vehicles often move with high
bandwidth will be wasted. On the contrary, if the speed. This implies that the time duration for
slots in a frame are fully occupied, a new vehicle exchanging data between a vehicle and RSU is
Figure 11. The scenario considered in RSU scheduler
193
Figure 12. Packets in D-list and S-list are sorted by Deadline and Datasize respectively in RSU
very short. As shown in Figure 12, when a large buffered in a queue of RSU at the beginning.
number of vehicles intend to exchange data with Each request is characterized by a 4-tuple:
RSU simultaneously, an efficient scheduler that < u - id, d - id, op, deadline > , where u - id
arranges a sequence of uplink and downlink data is the identifier of the vehicle, d - id is the identi-
access is important. How to decide an efficient fier of the requested data item, op is the operation
serving sequence for satisfying all the services that the vehicle wants to do (upload or download),
requested from vehicles is an imperative issue. and deadline is the critical time constraint of the
Since each vehicle can only exchanged data request. In addition to the deadline, the data size
with the RSU within a very short time duration should also be taken into consideration because
which highly depends on the vehicle speed and that a transmission for a large sized data needs long
the communication range of RSU, each request service time and may delay other service requests.
sent from a particular vehicle would have a time As shown in Figure 12, the RSU will initially sort
constraint. This means that the request would be all the requests according to the deadline and data
overdue if the request did not be served within the size individually and store the sequences as the
time constraint. However, the RSU might receive a D-list and S-list, respectively.
number of requests with different time constraint. A weighted DS_value, as defined in Equation
Since the bandwidth resource is limited, how to 1, is used for RSU to decide the services schedul-
maximize the network throughput and prevent ing sequence.
the requests from being overdue will be the major
goal for developing an RSU scheduler. DS _ value = (Deadline - CurrentClock ) ´ DataSize
Zhang et al. (2007) proposed a RSU sched- (1)
uling mechanism that satisfies the services
requested from many vehicles under the time
The (Deadline - CurrentClock ) represents
and bandwidth constraints. As shown in Figure
the remaining service time. A service with a smaller
12, assume that all of the service requests are
194
time constraint will have a higher priority while a between nodes i and j, and R denote the average
service with a smaller DataSize will have a higher transmission range. The forwarding probability,
priority. Consequently, the DS scheme always pij, can be calculated on a per packet basis using
serves the requests with minimum DS _ value the following simple Equation. The basic concept
in each scheduling time. of Equation 2 is that a receiver closer to the sender
will have a larger overlapped communication
range with sender and thus has smaller contribu-
emergency mac tion for forwarding the packet to those receivers
proTocols on VaneTs that have not yet received the same message
from the sender. As a result, a receiver that farer
broadcast strom mitigation to the sender will have a higher probability for
Techniques in VaneTs rebroadcasting the received message.
In VANETs, Broadcasting for emergency message Dij

has been widely discussed for a variety of appli- pij = (2)
R
cations. In the application of emergent message
dissemination, the packet broadcasting is critical To avoid the occurrence of contention, the
in time constraint. However, packet broadcasting receiver with small probability value should ad-
might raise contention and collision phenomenon ditionally wait for a predefined time duration. In
which might block the dissemination of emergency case of receiving duplicate packets from multiple
message. How to develop an efficient broadcast- sources within the waiting period, it calculates
ing mechanism is an important research issue in a probability for each received message and
VANETs. then selects the smallest probability value as its
Developing emergency MAC protocols can re-forwarding probability. If a node decides not
help dissemination of emergency messages with to rebroadcast, it should buffer the message for
a short propagation delay without packet colli- another time duration WAITTIME + d , where
sion and contention. Wisitpongphan et al. (2007) d is the one-hop transmission and propagation
assumed that each vehicle is aware of location delay, which is typically less than WAITTIME as
information and proposed three different broad- shown in Figure 13.
casting approaches. In the first approach, a node The proposed approach prevents the packet
j receiving a packet from node i will check the dissemination from collision. However, the first
packet ID and rebroadcasts with probability pij approach exhibits two drawbacks as described
if the packet is received first time. Otherwise, it below. First, the vehicle with low rebroadcast
discards the packet. Let Dij denote the distance priority may still broadcast the packets. Second,
Figure 13. Vehicles calculate probability and broadcast
195
if the density of VANET is high, several vehicles size and the larger the number of slots), it is very
may have the same probability. They will re- difficult for each vehicle to predict what the traf-
broadcast the packets at the same time, resulting fic density is and to arrive at a single value of
in collisions. The second approach, based on ê
Sij = N 8 - ê
(
é min RSS
range ( )
, RSSij - RSS min ) ´ N 8 ùú
ú
time-slot calculation, is proposed to improve the ê ú
RSSrange
aforementioned drawbacks. Upon receiving a êë úû
packet, each receiver checks the packet ID and in practice. Hence, network designers can, at best,
rebroadcasts with probability 1 at the assigned fix this value or adaptively change this value over
time slot TSij if the packet is received first time. time; for example the protocol should use five slots
Otherwise, it discards the packet. during morning and evening rush hour, and three
The following discusses how to derive the slots during non-rush hours. The second approach
value of the allocated transmission slot TSij. Let might assign the same slot for two vehicles that
the average transmission range and the predeter- have similar distances far from the sender. As a
mined number of slots are R and Ns, respectively. result, a collision might be occurred. Another ap-
Given the relative distance Dij between nodes i proach is proposed to improve this drawback.
and j, TSij can be calculated by Equation 4, where The proposed third approach takes into account
τ is the estimated one-hop delay, which includes the previously discussed two approaches. That is,
the medium access delay and propagation delay, probability-based and slot-based mechanisms are
and Sij is the assigned slot number. adopted in this approach. Each node in this scheme
should also buffer the message for a certain period
éN - 1ù ´WAIT + dms (3) of time (e.g., éëêN 8 - 1ùûú ´WAITTIME + dms ) and
êë 8 úû TIME
retransmits with probability 1 if nobody in the

TS = Sij ´ t (4) neighborhood rebroadcasts in order to prevent
ij
the message’s dying out.

The time slot approach follows the same The abovementioned approaches mainly devel-
concept of the weighted p-persistence scheme, oped based on the location information which is
but instead of calculating the re-forwarding obtained from GPS equipment. However, vehicles
probability, each node uses the GPS information may not be able to receive GPS signals in some
to calculate the waiting time to retransmit. For areas (e.g., tunnels, shadowed areas, urban areas
example, in Figure 13 the broadcast coverage with many high-rise buildings). These broadcast-
is spatially divided into four regions. According ing approaches can also be modified by using the
to Equation 4, a shorter waiting time will be as- packet Received Signal Strength (RSS) informa-
signed to the nodes located in the farthest region. tion instead of GPS information. Equations 5 and
When a node receives duplicate packets from 6 depict the modifications of approaches 1 and 2,
more than one sender, it takes on the smallest Dij where RSSrange and RSSmin denote the signal range
value. Similar to the p-persistence scheme, this and the value of minimum signal strength, respec-
approach requires transmission range information tively, and the RSSij denotes the signal strength
in order to agree on a certain value of slot size or between nodes i and j.
number of slots.
Note that N 8 is a designed parameter that RSSij - RSS min
should be carefully chosen. Although N 8 should pij = (5)
RSS range
theoretically be a function of the traffic density
(i.e., the denser the traffic, the smaller the slot
196
ê
Sij = N 8 - ê
(
é min RSS
range ( )
, RSSij - RSS min ) ´ N 8 ùú
ú
be broadcasted to those vehicles located at the
ê ú
forward location of the event. For example, the
RSSrange
ëê úû event message of ambulance alarming should be
(6) disseminated to the vehicles in the forward direc-
tion. Therefore, the broadcast areas of these two
types of events should be different.
broadcast methods for inter- Based on the concept, the authors proposed
Vehicle communications system a broadcast scheme to notify the vehicles in a
particular area. Figure 14 shows an example for
Some other studies devoted themselves to develop an ambulance alarming event. The authors define
broadcast schemes for emergency events such as three different areas in the road. The first area is
traffic accidents and ambulance alarming. The ac- called location offset, which avoids the inaccurate
cidents information should be broadcasted to all location obtained from a GPS. The second area is
vehicles to prevent them from another accident. used for relaying the emergency message along the
Therefore, broadcasting mechanism for emergen- forward direction. Vehicles in this area should not
cy information is important and has received much only receive the emergency notice but also deliver
attention recently. How to efficiently broadcast the the notice to other vehicles. Finally, the authors
emergency message to all vehicles neighboring to define the area which should be notified. In this
the accident location will be the key technology area, the vehicles within a proper range can get the
for preventing secondary accidents. notification by other vehicles without forwarding.
Fukuhara et al. (2005) proposed a novel broad- The following uses an example shown in Figure
cast method for a vehicular network to timely 15 to discuss how to identify the three areas.
notify appreciate drivers. They considered that Figure 15 depicts the area which consists of
the events can be categorized into two types, ac- an available relay range and an available notifi-

cording to the broadcasting directions. One type cation range. Let X and V be vehicle’s location
of events is backward-dissemination event which and velocity vectors, respectively. Equation 7(a)
should be broadcast to those vehicles located at depicts the relative direction between the vehicle
 
the backward location of the accident. A typical and the ambulance, where X 1 and X 2 denote
example is the traffic collision event. On the other the location vectors of vehicle and ambulance,
hand, the forward-dissemination event should
Figure 14. The relay and notification zone
197
Figure 15. An example for calculating available relay range
 
respectively, and V 1 and V 2 denote the velocity al. intend to reduce the dissemination packets in
vectors of vehicle and ambulance respectively. A a broadcasting operation while achieves high suc-
positive value of Equation 7(a) indicates that the cess rate. In their design, the packet only consists
vehicle and ambulance have the same direction. of the position of the accident, sender’s position,
Vehicles that satisfy the criteria of Equation 7(b) communication range and the propagation func-
should be notified and further relay the emergency tion including the target zone. In particular, the
message while vehicles satisfy Equation 7(c) but propagation function encodes the information
not satisfy Equation 7(b) should only be notified. about target areas and preferred routes.
The calculation procedure in the traffic accidents Figure 16 shows a propagation function as-
scenario is similar except the direction is opposite sociated with a target zone that is reached by a
to the ambulance alarming event. single major road. The function drives messages
along the main road — the red line below the
    function — and towards the target zone — the
( )
X 2 - X 1 + aV 1 ·V 2 > 0 (7a)
black ellipse. It is worth to note that the message
originator does not compute a predefined trajec-
dis tan ce (m.SenderPosition, localPosition )
p¬ tory using the propagation function before sending
m.SenderCommunicationRadius the message. Rather, the route to the destination
(7b) is the result of the evaluation of the function at
   each routing hop. For example, a message that is
X 2 - X 1 + aV 1 < RNotification (Threshold) (7c) routed outside the black line in Figure 16 does not
have to be routed back towards the line, but it can
continue its route along a new trajectory, which
Towards lightweight still ends up on the desired target area. The values
information dissemination in of the propagation function can be viewed as the
inter-Vehicular networks values of the potential to achieve the destination
region. Messages should be attracted in the right
In addition to the traffic accidents and ambulance direction towards decreasing values of the propa-
alarming handling, some other emergency proto- gation function that is towards areas of minimum
cols are developed for general-type emergency potential. Several protocols with the propagation
event (Sormani et al., 2006). Davide Sormani et
198
Figure 16. Three-dimension diagram of the

and target area is shorter than that between sender
propagation function
and target area, the function value calculated by
the receiver is smaller than that calculated by the
sender. That is, vehicle C has the smaller function
value than vehicle B and vehicle A has the larger
function value than vehicle B. As a result, vehicle
C in the OZF protocol gets the value “one” to relay
the packet and vehicle A gets the value “zero”
which indicates it does not need to forward the
received packet. The OZF mechanism can reduce
the phenomenon of packet dissemination along the
opposite direction toward the destination. How-
ever, several vehicles might forward the received
message at the same time if they are within the
transmission range of a sender and are closed to
the destination. This might raise the contention
and collision problem.
b. Distance-Driven Probabilistic Diffusion
function are presented for broadcasting the emer- (DDPD). DDPD is proposed for reducing the
gency message with lightweight traffic. collision and contention phenomenon occurred
a. One Zero Flooding (OZF). The One Zero in OZF. The DDPD propagation protocol adds a
Flooding (OZF) mechanism is proposed for flood- probabilistic choice to OZF. The probability func-
ing a message with lightweight traffic by using tion introduced in DDPD design aims to distribute
propagation function. The basic concept of OZF those transmissions which might be collided with
mechanism is that packets would be transferred each other when applying OZF. Equation 8 takes
by vehicles with the smaller value of propagation into account the distance between the receivers
function. A simple example to illustrate the idea and the sender. Similar to the concept of OZF,
of OZF is shown in Figure 17. In Figure 17, we vehicles do not forward the emergent messages if
assume that vehicles A and C both are located these vehicles have larger values of propagation
within vehicle B’s communication range. We as- function than sender. On the contrary, vehicles
sume that vehicle A and vehicle C calculate the with smaller values of propagation function than
function value between themselves and vehicle sender will forward the emergent messages to the
B separately after getting a packet broadcasted target area with a probability evaluated according
from vehicle B. If the distance between receiver to Equation 8.
Figure 17. OZF propagation protocol
199
dist (X ,Y ) + Du * BB _ REFR would be a better candidate for broadcasting the

FF (X ) = (8) message than B since it closer to the target area.
R
c. Function-Driven Probabilistic Diffusion
Figure 18 depicts an example of DDPD. As- (FDPD). Though the abovementioned DDPD
sume that vehicles B and C have smaller function mechanism tries to reduce the number of forward-
values than vehicle A and intend to forward the ers, however, it does not consider the distance
received message with a probability calculated from forwarding candidates to destination in the
by Equation 8. Herein, we assume that vehicle C probability design. To further reduce the number
has a higher probability value than vehicle B and of redundant transmission, FDPD propagation
will forward the received emergency message. protocol is proposed. The probability function
As a result, vehicle B will not further forward proposed in FDPD relies on the function value
the message again after it overhears the message by preventing the packet from forwarding in the
broadcasted from vehicular C. Although the wrong direction. Different to DDPD, an enhance
DDPD mechanism might reduce the probability probabilistic Equation is proposed as shown
of transmission collision, it still has the problem in Equation 9, where the BEST point denotes
of redundant transmission. For example, in Figure the physical location with the lowest value of
19, if vehicles B and C are closed to the edge of propagation function within the communication
the communication range of vehicle A, vehicle radius of the sender. In the FDPD propagation
B will have a high probability for forwarding the protocol, vehicles forward the emergent packet if
packet sent from vehicle A. However, vehicle C they are located at the right position away from
Figure 18. DDPD propagation protocol
Figure 19. Vehicles B and C have the same high probability for forwarding the packet sent from vehicle
A. However, vehicle C would be a better candidate
200
the BestPoint and they are closer to the destina- FSFD mainly adopts the well known store-and-
tion than sender. forward mechanism where vehicles can store the
emergent packet and move for a period of time,
f (m.SenderPosition ) - f (localPosition ) seeking for the possible forwarder. As shown in
p¬ Figure 21, assume that vehicle B can not find
f (m.SenderPosition ) - f (BestPo int)
any forward in its transmission range at the time
(9) t after receiving the packet from vehicle A. Thus,
vehicle B has to store the packet in its queue
As shown in Figure 20, vehicle B and vehicle temporally until it finds forwarder C at the time
C intend to forward the received packet with t + 2. In the meanwhile, vehicle B forwards the
probability 0.75 and 0.99, respectively, according packet to vehicle C and therefore the packet will
to Equation 9. The key improvement of FDPD not be blocked at vehicle B.
is the consideration of BEST point which leads e. Direction-aware Function Driven Feedback-
the packet to be forwarded in the right direction. augment Store & Forward Diffusion (DFD-
Though FDPD significantly reduce the number FSFD)
of redundant transmissions, however, it does not Although the FSFD can store and forward the
take into consideration the hole problem, which packet for overcoming the hole problem, however,
might occur in a sparse network environment. the transmission direction of packet is not taken
d. Feedback-Augmented Store and Forward into account. This protocol takes into account the
Diffusion (FSFD). Compared with the FDPD, the right direction to target area like the BEST point
FSFD further concerns the hole problem which and applies the FSFD propagation protocol to
might occur in the light-traffic environment. The forward the emergent packet. As shown in Figure
Figure 20. FDPD propagation protocol
Figure 21. FSFD propagation protocol
201
Figure 22. DFD-FSFD propagation protocol
22, the vehicles A and B calculate the angles α and those nodes in the backbone topology. This sub-
p section introduces a novel reliable MAC protocol
β, respectively. The angle α less than implies which efficiently constructs a backbone topology
2
that the packet can be forwarded by vehicle A for a given VANET.
to the right direction to the target area. On the Bononi et al. (2007) proposed a fast and effi-
p cient broadcast protocol for VANET. The proposed
contrary, the angle β larger than indicates
2 protocol aims to reduce the influence of vehicles
that the packet forwarded by vehicle B is along
with high speed on performance. The following
the wrong way.
uses an example to illustrate this protocol.
As shown in Figure 23, a distributed clustering
dynamic algorithm is proposed to create a dynamic
reliable mac proTocols
virtual backbone in the vehicular network. The
on VaneTs
vehicles update their state dynamically. If the
vehicles are in the backbone, their states are set
Some other studies took the reliable communica-
to be Backbone Member (BM), otherwise their
tion into consideration and aimed to construct a
states are set to be Normal Vehicle (NV). The
reliable network topology for data dissemination.
vehicles in BM are responsible for relaying the
However, in VANETs, the locations of vehicles
massages, while the vehicles in the NV state just
change with time, which might cause an existed
need to receive the messages from the vehicles in
link broken. It is an important issue to construct
the BM. The proposed algorithm can reduce the
a reliable backbone topology so that the message
overhead of inter-vehicular communication.
flooding can be achieved by the forwarding of
Figure 23. The vehicles in BM are responsible for relaying the massages, while the vehicles in the NV
state just need to receive the messages from the vehicles in the BM
202
A backbone creation process starts whenever the average speed of vehicles X and Y, respectively.
a vehicle does not receive backbone beacons for In Equation 10 gives the estimation of residual
a predefined time interval. As shown in Figure time of vehicles X and Y, where sign() denotes
24, vehicles A, B, C and D initially implement the function which returns 1 if Du is positive
the random back-off mechanism without receiv- and -1 otherwise.
ing a message from backbone. In this example,
we assume that vehicle A wakes up earliest and ( )
é max 0, sign Du ù R - dist X ,Y
ê ( ) úû ( )
sets itself as a backbone member. Then vehicle RT (X ,Y ) = ë
Du
A chooses its next hop backbone member from
neighbors for constructing a robust backbone, (10)
the next hop should be able to stay in the com-
munication range of A for the longest time. As In order to reduce the communication overhead
a result, the backbone member can be selected required for constructing a backbone, the number
subsequently in a hop-by-hop manner. Each ve- of vehicle in BM state should be minimized. To
hicle in the BM state should record at most two achieve this goal, Equation 11 calculates the dis-
backbone members which are called previous hop tance of X and Y after a time duration BB_PEFR
and next hop, respectively. Figure 25 shows that which is the time required for constructing the
vehicles B and C are the previous hop and next next backbone topology, normalized with respect
hop of vehicle A, respectively. to R.
An important factor for a backbone member
to select the next backbone member will be their dist (X ,Y ) + Du * BB _ REFR
FF (X ) =
connection duration. The following Equation is R
used to calculate the residual communication (11)
time between vehicles X and Y. Let RT denote
the Residual Time, R denotes the communication If the distance between two vehicles increases,
range of sender vehicle, and dist(X, Y) denotes the value of FF would raise. The value of FF
the current estimated distance of vehicles X and would be used as the random backoff parameter
Y. Let Du = uY - uX denotes the relative speed after all neighbors of vehicle A calculating RT.
between vehicles X and Y, where uX and uY are Therefore, when a vehicle has the largest value
it will wake up first and response the beacon
Figure 24. Vehicle A, B, C and D do not receive backbone beacons for a time interval. After random
backoff time period, if the first waked vehicle is A, vehicle A will elect itself as a backbone member and
choose its next hop.
203
Figure 25. Each vehicle in the BM state has at most two neighbors that called previous hop and next
hop respectively. Previous hop and next hop are also backbone members.
of vehicle A. We assume that vehicle C wake protocols that meet the requirement of real applica-
up earliest and it responds a packet to A. Upon tions in VANETs will be the future works.
receiving the response, Vehicle A records that C
is its next hop, and then sends an ACK_WINNER
packet to vehicle C. Similarly, vehicle C records reFerences
that vehicle A is its previous hop and changes the
NV state to BM state. All other vehicles give up ASTM International E2213-03 (2003). Standard
the contention accordingly. Specification for Telecommunications and In-
formation Exchange Between Road-side and
Vehicle Systems 5 GHz Band Dedicated Short
conclusion Range Communications (DSRC) Medium Ac-
cess Control (MAC) and Physical Layer (PHY)
Vehicular communication systems that can trans- Specifications.
mit and receive information to and from individual Bononi, L., & Felice, M. D. (2007). A Cross Lay-
vehicles have the potential to significantly increase ered MAC and Clustering Scheme for Efficient
the safety of vehicular transportation, improve Broadcast in VANETs. In Proceedings of the IEEE
traffic flow on congested roads, and decrease International Conference on Mobile Ad hoc and
the number of people of deaths and injuries in Sensor Systems.
vehicular collisions effectively.
This chapter review import MAC protocols for Fukuhara, T., Warabino, T., Ohseki, T., Saito, K.,
VANETs. A number of MAC protocols, including Sugiyama, K., Nishida, T., & Eguchi, K. (2005).
IEEE 802.11p, DSRC standard, TDMA-based Broadcast Methods for Inter-Vehicle Communica-
MAC protocols, data access MAC scheduling tions System. In Proceedings of the IEEE Wireless
protocol, emergency message and reliable MAC Communications and Networking Conference.
protocols, are reviewed. In this chapter, the basic
IEEEP802.11p (2009, May). IEEE Draft Standard
concept, strategies, advantages and contribu-
for Information Technology – Telecommunica-
tions of the reviewed papers are introduced and
tions and information exchange between systems
discussed. We believe the reviewed MAC proto-
– Local and metropolitan area networks – Specific
cols will be the base for further improvement of
requirements – Part 11: Wireless LAN Medium
protocols in terms of the reliability, scalability,
Access Control (MAC) and Physical Layer (PHY)
and efficiency in VANETs. The development of
specifications Amendment 7: Wireless Access in
efficient VANET protocols that integrate MAC
Vehicular Environments.
and network layers and the implementation of
204
Draft 7.0, IEEE Standard 1609.1 (2006). IEEE Tri- IEEE Standard 1609.3 (2007). IEEE Trial-Use
al-Use Standard for Wireless Access in Vehicular Standard for Wireless Access in Vehicular Envi-
Environments (WAVE) – Resource Manager. ronments (WAVE) – Networking Services.
Korkmaz, G., Ekici, E., & Ozguner, F. (2006). Wisitpongphan, N., Tonguz, O. K., Parikh, J. S.,
A Cross-Layer Multihop Data Delivery Protocol Mudalige, P., Bai, F., & Sadekar, V. (2007). Broad-
with Fairness Guarantees for Vehicular Networks. cast Strom Mitigation Techniques in Vehicular Ad
IEEE Transactions on Vehicular Technology, Hoc Networks. IEEE Wireless Communications,
55(3), 865–875. doi:10.1109/TVT.2006.873838 14(6), 84–94. doi:10.1109/MWC.2007.4407231
Sormani, D., Turconi, G., Costa, P., Frey, D., Yu, F., & Biswas, S. (2007). Self-Configuring
Migliavacca, M., & Mottola, L. (2006). To- TDMA Protocols for Enhancing Vehicle Safety
wards Lightweight Information Dissemination with DSRC Based Vehicle-to-Vehicle Commu-
in Inter-Vehicular Networks. In Proceedings of nications. IEEE Journal on Selected Areas in
the ACM International Workshop on Vehicular Communications, 25(8), 1526–1537. doi:10.1109/
Inter-Networking. JSAC.2007.071004
IEEEStandard1609.2 (2006). IEEE Trial-Use Zhang, Y., Zhao, J., & Cao, G. (2007). On Schedul-
Standard for Wireless Access in Vehicular En- ing Vehicle-Roadside Data Access. In Proceedings
vironments (WAVE) – Security Services for Ap- of the ACM International Workshop on Vehicular
plications and Management Messages. Inter-Networking.
IEEE Standard 1609.4 (2006). IEEE Trial-Use
Standard for Wireless Access in Vehicular Environ-
ments (WAVE) – Multi-channel Operation.
205
206
Chapter 13
Routing Protocols in
Vehicular Ad Hoc Networks
Yuh-Shyan Chen
Yun-Wei Lin
absTracT
Vehicular Ad hoc Network (VANET), a subclass of mobile ad hoc networks (MANETs), is a promising
approach for the intelligent transportation system (ITS). The design of routing protocols in VANETs is
important and necessary issue for support the smart ITS. The key difference of VANET and MANET is
the special mobility pattern and rapidly changeable topology. It is not effectively applied the existing
routing protocols of MANETs into VANETs. In this chapter, we mainly survey new routing results in
VANET. The authors introduce unicast protocol, multicast protocol, geocast protocol, mobicast proto-
col, and broadcast protocol. It is observed that carry-and-forward is the new and key consideration for
designing all routing protocols in VANETs. With the consideration of multi-hop forwarding and carry-
and-forward techniques, min-delay and delay-bounded routing protocols for VANETs are discussed in
VANETs. Besides, the temporary network fragmentation problem and the broadcast storm problem are
further considered for designing routing protocols in VANETs. The temporary network fragmentation
problem caused by rapidly changeable topology influence on the performance of data transmissions.
The broadcast storm problem seriously affects the successful rate of message delivery in VANETs. The
key challenge is to overcome these problems to provide routing protocols with the low communication
delay, the low communication overhead, and the low time complexity.
inTroducTion municate with other vehicles to form a special

class of wireless networks, known as vehicular ad
The growth of the increased number of vehicles hoc networks or VANETs (Saha et al., 2004; Xu
are equipped with wireless transceivers to com- et al., 2004; Yousefi et al., 2004). Vehicular Ad
hoc Network (VANET) is a promising approach
DOI: 10.4018/978-1-60566-840-6.ch013 for the intelligent transportation system (ITS)
Routing Protocols in Vehicular Ad Hoc Networks
(ASTM E2213-03, 2003). Vehicle information divided into three cases: (1) a static destination or a
is delivered via the multi-hop wireless transmis- mobile destined vehicle, (2) more than one vehicle
sion over VANETs to provide safety or comfort in a geographic region, and (3) all vehicles within
applications for drivers. VANETs are expected to a geographic region. When the message should be
improve the traffic quality and provide the more sent to a static destination or a mobile destined ve-
convenient driving environment for the general hicle, unicast routing protocol is utilized. Unicast
populace. It is known that VANET (Saha et al., routing is a fundamental operation for vehicle to
2004; Xu et al., 2004; Yousefi et al., 2004) is a construct a source-to-destination path in a VANET.
subclass of mobile ad hoc networks (MANETs) Example of unicast routing is given in Figure
(Briesemeister et al., 2000) . Just like a MANET, 1(a). For the case of the static destination, the
a VANET has no the fixed infrastructure. In ad- routing path is quite different because the source
dition, vehicles with high speed mobility make vehicle is continuously moving at the different
VANETs having quite different characteristics time period. It is more difficult to continuously
from MANETs; such as rapidly changed topology trace the location of destination vehicle if the
and frequently network fragmentation. VANETs destination is a mobile vehicle. Moreover, when
are mainly realized in city and highway environ- the information is delivered to more than one ve-
ments. Roads and streets with intersections is the hicle in a geographic region, multicast and geocast
major scenario in the city environment. Multiple routing protocols are performed. For the geocast
lanes with single or dual direction are investigated routing, if a vehicle receives a geocast packet
in the highway environment. These two environ- from neighbors, the packet should be forwarded
ments have different impact on VANETs. In the or dropped depended on its current location. If
city environment, packets are difficult to be suc- this vehicle is located in the specific geographic
cessfully transmitted since the signals are easily region, the geocast packet is forwarded; otherwise,
shielded by buildings. With the obstacles, two the packet is dropped. Multicast in a VANET is
vehicles are not able to be communicated even if defined by delivering multicast packets from a
they are very close. In the highway environment, single source vehicle to all members of multicast
the temporary network fragmentation problem is in a multi-hop communications as shown in Figure
the key issue. Furthermore, VANETs have distinc- 1(b). The multicast and geocast routing protocol
tive features. For example, power constraint is also are important functions for many useful ap-
not the major concern, and location information plications, including collision warning system,
is easily obtained from GPS (Global Positioning distributed games, replicated file systems, and
System) (Gerten et al., 2005) which is the com- teleconferencing. Broadcast protocol is utilized
mon equipment in a vehicle. The interest problem if the information should be sent to all vehicles
is how to develop the efficient routing protocols in a network. A source vehicle sends broadcast
in VANETs with the consideration of distinctive message to all other vehicles in the network.
features of VANETs. Example is illustrated in Figure 1(c). Broadcast
The design of routing protocols of VANETs is an important function in many applications
is the important issue for supporting the smart of VANETs, such as advertisement publicity,
ITS. To enhance the safety of drivers and provide cooperative operations, group discussions, and
the comfortable driving environment, messages route discovery. The design issue of broadcast-
for different purposes need to be sent to vehicles ing is how to effectively prevent packet collision
through the inter-vehicle communications. Ac- and reduce the broadcasting overhead during the
cording to the number of receiving vehicles within broadcasting.
a geographic region, the roles of destinations are
207
Figure 1. Applications for VANET (a) unicast routing, (b) geocast and multicast routing (c) broadcast
routing
Many results (Safa et al., 2009; Sawamura et bounded routing protocol (Skordylis et al., 2008) is
al., 2008; Manoharan et al., 2008; Hughes et al., to establish a minimum communication overhead
2006; Manoharan et al., 2005) on MANETs have routing path under a constrained delay time. For
been proposed for unicast, multicast and geocast, the min-delay routing protocol, we review GPCR
and broadcast protocols. However, VANETs are (Lochert et al., 2005), VADD (Zhao et al., 2006),
fundamentally different to MANETs, such as the CAR (Naumov et al., 2007), DIR (Chen et al.,
property of mobility and rapid changed topology 2009), ROMSGP (Taleb et al., 2007), reliable
(Blum et al., 2004); therefore, VANET is subject routing (Wan et al., 2008), and GVGRID (Sun et
to frequently network fragmentation and a stable al., 2006). For the delay-bounded routing protocol,
routing path is not easily to establish. This key we describe delay-bounded routing in (Skordylis
differentiation causes the existing routing proto- et al., 2008). In multicast and geocast approaches,
col on MANETs can not be directly applied to two categories of protocol are discussed, spatial
VANETs. In this investigation, the recent new routing protocol and spatiotemporary routing
results for VANET routing mechanism are first protocol. For spatial routing protocol, DRG (Joshi
surveyed. As shown in Figure 2, the survey is et al., 2007) and IVG (Bachir et al., 2003) are
structured into three broad categories; unicast, discussed. For spatiotemporary routing protocol,
multicast and geocast, and broadcast approaches. a mobicast routing protocol (Chen et al., 2009)
It is observed that carry-and-forward is the new is introduced. Finally, the broadcast approaches
and key consideration for designing all routing are investigated. The impact of broadcast storm
protocols in VANETs. The key ideas of representa- problem on MANETs has been firstly defined
tive technologies in each category are described. in (Tseng et al., 2002), then some researchers
In the unicast approaches, existing unicast routing (Tonguz et al., 2006) further discusses the im-
protocols are classified into min-delay and delay- pact of broadcast storm problem on VANETs.
bounded routing protocols. The min-delay routing Besides, two broadcast approaches for VANETs,
protocols (Lochert et al., 2005; Zhao et al., 2006; DV-CAST (Tonguz et al., 2007) and broadcast
Naumov et al., 2007; Chen et al., 2009; Taleb et methods for V2V communications (Fukuhara et
al., 2007; Sun et al., 2006; Wan et al., 2008) are to al., 2005), are finally reviewed. The challenges
construct a routing path with the minimum delay and perspectives of routing protocols for VANETs
time for a source-to-destination path. The delay- are finally discussed.
208
The remainder of this paper is organized as fol- is normally longer than using wireless multi-hop
lows. Section 13.2 reviews unicast routing proto- transmission technique. Two categories of routing
cols in VANETs. Section 13.3 introduces multicast protocol designing are classified, min-delay rout-
and geocast routing protocols in VANETs. Section ing protocol and delay-bounded routing protocol.
13.4 describes broadcast routing protocols in Min-delay routing protocol aims to minimize the
VANETs. Section 13.5 concludes this chapter. delivery delay-time from source to destination.
Delay-bounded routing protocol attempts to main-
tain a low degree of channel utilization within the
unicasT rouTing proTocol constrained delivery delay-time. Figure 3 gives
the classification of these unicast protocols.
This section introduces the unicast routing proto-
cols in VANETs. The main goal of unicast routing min-delay routing protocol
in VANETs is to transmit data from a single source
to a single destination via wireless multi-hop The goal of min-delay routing protocols is to
transmission or carry-and-forward techniques. In transmit data packets to destination as soon as
the wireless multi-hop transmission, or called as possible. The transmission delay time is the major
multi-hop forwarding, technique, the intermedi- concern and the shortest routing path is usually
ate vehicles in a routing path should relay data as adopted. However, the shortest routing path may
soon as possible from source to destination. In the be not the quickest path with the minimum delay
carry-and-forward technique, source vehicle car- time in VANETs. The shortest routing path may
ries data as long as possible to reduce the number be found in a low density area, packets can not
of data packets in the VANETs. The delivery delay- transmit by the multi-hop forwarding since that
time cost by using carry-and-forward technique there is no neighboring vehicle can forward
Figure 2. The taxonomy of vehicular ad hoc networks
209
Figure 3. Unicast routing protocols
packets. Packets should be delivered by carry- packets. Forwarding between junctions is then
and-forward scheme. The delay time is greatly done in a position-based fashion. Observe that,
growing if the multi-hop forwarding can not be GPCR does not need the global knowledge of the
utilized. Efforts will be made as finding a routing street map. GPCR traverses the junctions by a re-
path with multi-hop forwarding. The min-delay stricted greedy forwarding procedure, and adjusts
routing protocols (Lochert et al., 2005; Zhao et the routing path by the repair strategy which is
al., 2006; Naumov et al., 2007; Chen et al., 2009; based on the topology of streets and junctions.
Taleb et al., 2007; Sun et al., 2006; Wan et al., Figure 4 shows that vehicle Vu tries to send packets
2008) are reviewed as follows. to vehicle VD. Vehicle 1a is selected as the next
hop of Vu if greedy forwarding scheme is used.
Greedy Perimeter Coordinator After vehicle 1a received the packets, vehicle
Routing Protocol 1a detects destination VD is not located at north.
Vehicle 1a then moves packets backward vehicle
Lochert et al. (2005) proposed GPCR (greedy 2a, then the packet is forwarded to VD. It shows
perimeter coordinator routing) which is a position- that GPCR not require the global knowledge of
based routing for urban environment. The basic the city map.
idea just likes Greedy Perimeter Stateless Routing
(GPSR) (Karp et al., 2000). The routing path is
Figure 4. Geographic routing protocol
established by the greedy method. GPCR protocol
is very well suited for highly dynamic environ-
ments such as inter-vehicle communication on the
highway or city. However, urban environment has
the radio obstacles problem (Karp et al., 2001).
Radio obstacles problem has the influence on the
performance of position-based routing. Therefore,
GPCR overcomes this problem by establishing ro-
bust routes in the city environment. GPCR adopts
the similar idea of GPSR (Blaevi´c et al., 2002)
for the selection of intermediate nodes. GPSR
needs the global knowledge of the city topology
with the street map. The sender determines which
junctions have to be traversed by some control
210
VADD: Vehicle-Assisted Date and uij is the average vehicle velocity on rij. This
Delivery Routing Protocol delay model is very useful to estimate the data
delivery delay in VANETs.
Data delivery routing protocol is developed by With the delay model, VADD protocol esti-
Zhao et al. (2006). They proposed a vehicle- mates the best road with the lowest data delivery
assisted data delivery protocol, called VADD, in delay based on the current kept traffic patterns. In
VANETs. The VADD protocol adopted the idea fact, it is difficult to find a road with the minimum
of carry-and-forward for the data delivery. The forwarding delay between two arbitrary intersec-
most important issue is to select a forwarding tions from unlimited unknown intersections. In
path with the smallest packet delivery delay. It addition, observe that VADD only considers how
usually chooses the next hop closer to the desti- to find a path from a mobile vehicle to a fixed
nation for the data delivery, but this strategy is location of destination vehicle. However, it is
not suitable for the sparsely connected vehicular not easily collect the in time traffic pattern and
networks. In the sparsely connected vehicular information. From the out-of-date traffic infor-
networks, transmission loss commonly occurs mation, VADD may establish the inappropriate
in case of disconnection, the packet has to be road with the greater data delivery delay. Figure
carried by the vehicle, while the moving speed is 5 illustrates that vehicle Va tries to send packets
significantly slower than multi-hop forwarding. to the coffee shop, while the coffee shop is at the
To keep the low data transmission delay, VADD fixed location. Intersections Ia, Ib, Ic, and Id, are
protocol transmits packets through wireless chan- considered as the candidate intermediate intersec-
nels as much as possible, and if the packet has tions. After evaluating the expected forwarding
to be carried through roads, the road with higher delay, intersections Ia, Ic, and Id are chosen. This
speed is chosen firstly. VADD protocol assumes is because that the density of vehicle is high be-
that vehicles are equipped with pre-loaded digital tween intersections Ia, Ic, and Id, although it is not
maps, which provide street-level map and traf- the shortest path.
fic statistics such as traffic density and vehicle
speed on roads at different times of the day. Ac- Connectivity-Aware Routing Protocol
cording to the information provided by digital
maps, VADD protocol proposed a delay model To overcome the limitation of the fixed location of
to estimate the data delivery delay in different destination vehicle, Naumov et al. (2007) proposed
roads as follows, Connectivity-Aware Routing (CAR) protocol in
VANETs. The CAR protocol is a position-based
lij ´ c lij routing scheme. CAR protocol establishes a rout-
(
dij = 1 - e
-R´rij
)´ R
+e
-R´rij
´
uij
, ing path from source to destination by setting the
anchor points at intermediate intersections. Each
vehicle exchanges hello messages to collect the
where dij is the expected packet forwarding de- information of neighboring vehicle, such as mov-
lay from intersection Ii to intersection Ij, R is the ing direction and speed. CAR sends the search-
communication range of vehicle, c is a constant ing packets by using PGB algorithm (Preferred
used to adjust expected packet forwarding delay Group Broadcast) (Naumov et al., 2006) to find
to a more reasonable value, rij, is the road from the destination and a routing path from source to
intersection Ii to intersection Ij, ρij is the vehicle destination. Each forwarding vehicle records it’s
density on rij, lij is the Euclidean distance of rij, ID, hop count, and average number of neighbors
211
Figure 5. The environment for VADD routing protocol
in searching packets. Once the searching packets (DIR) protocol in VNETS. The key difference
reach the destination, the destination chooses a of CAR and DIR protocols is that DIR protocol
better routing path which has the minimum de- (Chen et al., 2009) constructs a series of diagonal
livery delay time and sends the reply packet to intersections between the source and destination
source. While destination sends the reply packet vehicles. The DIR protocol is a geographic rout-
to source, the intersections passed through by the ing protocol. Based on the geographic routing
reply packet are set as the anchor point. After the protocol, source vehicle geographically forwards
path set up, data packets are forwarded in a greedy the data packet toward the first diagonal intersec-
method toward the destination through the set of tion, the second diagonal intersection, and so on,
anchor points using the AGF algorithm (Advanced until the last diagonal intersection, and finally
Greedy Forwarding) (Naumov et al., 2006). To geographically reaches to the destination vehicle.
maintain the routing path, a guard node is used if For given a pair of neighboring diagonal intersec-
the destination changes the position. The guard tions, two or more disjoint sub-paths exist between
node indicates the routing path to destination. them. The novel property of DIR protocol is the
Figure 6 gives that vehicle VS tries to send data auto-adjustability; while the auto-adjustability is
to vehicle VD, the anchor points are set at I1,1, I2,1, achieved that one sub-path with low data packet
I2,2, I3,2 and I3,4 . Data is forwarded according to delay, between two neighboring diagonal inter-
order in the list of anchor points. sections, is dynamically selected to forward data
packets. To reduce the data packet delay, the route
DIR: Diagonal-Intersection- is automatically re-routed by the selected sub-path
Based Routing Protocol with lowest delay. Figure 7 shows that DIR pro-
tocol constructs a series of diagonal intersections
To improve the CAR protocol, Chen et al. (2009) between vehicles VS and VD. Observe that, DIR
developed a diagonal-intersection-based routing protocol may set the fewer number of anchors
212
Figure 6. CAR is an anchor-based routing protocol
than CAR protocol (Naumov et al., 2007). DIR vehicles receive packets from this sender vehicle.
protocol can automatically adjust routing path Figure 8 illustrates the ROMSGP routing protocol.
for keeping the lower packet delay, compared to Two routing paths are established, {VAVB, VBVD}
CAR protocol (Naumov et al., 2007). and {VAVC, VCVD}. If VA, VB, VC, and VD belong to
the same group, the two routing paths are both
ROMSGP Routing Protocol stable. Packet is delivered via {VAVB, VBVD} or
{VAVC, VCVD}. If VB turns into another road, the
To improve the routing reliability, Taleb et al. projection vector is changed. VB belongs to the
(2007) proposed ROMSGP (Receive on Most other group. Then the routing path {VAVC, VCVD}
Stable Group-Path) routing protocol in a city is the only choice.
environment. Taleb et al. (2007) indicate that an
unstable routing usually disconnects due to the Reliable Routing for Roadside
loss of connectivity. This dis-connectivity is oc- to Vehicle Communications
curred if one vehicle moves out of the transmission
range of a neighboring vehicle. A large difference In contrast with routing results developed in the
of velocity exists between a pair of two vehicles. highway or the city environments, it is very in-
In ROMSGP protocol, all vehicles are split into terest that Wan et al. (2008) specially proposed
four groups based on the velocity vector. A rout- a reliable routing protocol in the rural environ-
ing is said as a stable routing if the two vehicles ment. Wan et al. (2008) proposed two reliable
are categorized in the same group. If two vehicles routing strategies for roadside to vehicle (R2V)
are categorized in different groups, the routing is communication. The challenge of R2V commu-
an unstable routing. Each group has a unique ID nication in the rural environment is the terrain
and a base vector. A vehicle belongs to a group if factor. For instance, a vehicle moving along the
the velocity vector has the maximum projection rural highway occasionally loses the line of sight
vector with this group. The stability of routing (LOS) to the neighbor vehicle or to access points
is evaluated by comparing the group ID if other (APs) due to the obstacle-property caused by
213
Figure 7. DIR protocol sets fewer anchors than CAR protocol
Figure 8. ROMSGP routing protocol
the curve roadway and mountains. In addition, munication is established, the link lifetime halts
almost no fixed communication infrastructure is if (1) LOS between a pair of vehicles is lost, or
available. Multi-hop inter-vehicle communica- (2) one vehicle moves out of the communication
tion connecting to AP is the main solution of the range of the neighboring vehicle.
R2V communication. High mobility causes the A link established in a shorter distance usu-
temporary network fragmentation problem. The ally has longer link lifetime. The link lifetime is
link lifetime is very important issue for designing used to predict the lifetime of a route. A route is
the reliable routing to overcome the temporary constructed by a series of links. The lifetime of a
network fragmentation problem. The link lifetime route is the minimum link lifetime in a route. Long
is predicted by two conditions. Once the com- lifetime of a route improves the routing reliability
214
if considered the lifetime-bounded shortest path. routing path between roadsides and vehicles, or
In addition to the lifetime of a routing path, the between vehicles. GVGrid assumes each vehicle
length-bounded maximum lifetime path is con- has digital map and location information. GVGrid
sidered. To construct a length-bounded maximum divides the map into several grids. The RREQ and
lifetime path, reducing hops can improve the RREP packets are delivered through different grid
delivery delay-time. A routing path with fewer to find a routing path through minimum number
hops means the links are established in the long of grid. A grid is chosen based on the direction
distance. Establishing a routing path with longer and the distance between vehicle and intersection
lifetime implies that the length of this routing and is selected as next grid if the direction of grid
path is long. is the same as current grid or the grid is closed
Figure 9 (a) illustrates the example of lifetime- to the intersection. Then the intermediate grids
bounded shortest path. The dotted line is current between source and destination are recorded in
routing path and the link lifetime is going to the routing table. An appropriate vehicle which
end, where the minimum link lifetime is 9. The has the fewest number of disconnections in each
solid line is the candidate path. The link lifetime grid is chosen to forward packets to next grid. A
of solid line is greater than the threshold (=16). formula of evaluating the expected number of
The routing path changes to solid line by AP disconnections is derived in (Sun et al., 2006).
assignment. Figure 9(b) illustrates the example The routing table records in terms of the source
of length-bounded maximum lifetime path. The vehicle, destined grid, an appropriate vehicle as
dotted line is the routing path with minimum hops next hop with minimum the expected number
to AP (hops=4). The solid line is the selected path of disconnections, a vehicle as previous hop,
(hops=5). and the grid sequence. Once the routing path is
broken, GVGrid just finds another vehicle in the
GVGRID: A QoS Routing Protocol grid instead of the previous vehicle. The routing
path does not require finding again. Figure 10 (a)
To improve delivery delay-time and routing reli- shows that vehicle VS floods RREQ message to find
ability, Sun et al. proposed GVGrid protocol (Sun vehicle VD and vehicle VD replies RREP message
et al., 2006) which is a QoS routing protocol for to VS. Figure 10 (b) demonstrates that not only the
VANETs. GVGrid constructs a routing path from grid sequence but also the information of the next
source to destination by grid-based approach. vehicle are recorded in the routing table.
The goal of GVGrid is to maintain a high quality
Figure 9. (a) lifetime-bounded shortest path (b) length-bounded maximum lifetime path
215
Figure 10.(a) An example of GVGrid routing protocol, (b) routing tables on vehicle VA and VD
delay-bounded routing protocol 2008) focuses on the development of carry-and-

forward schemes that attempts to deliver data
Delay-bounded routing provides the low over- from vehicles to fixed infrastructure access point
head in wireless communication. Delay-bounded in an urban environment. Two routing algorithms,
routing delivers packet using carry-and-forward D-Greedy (Delay-bounded Greedy Forwarding)
method as much as possible. Delay-bounded rout- and D-MinCost (Delay-bounded Min-Cost For-
ing usually applies to non-real time services, such warding), evaluate traffic information and the
as FTP, e-mail, and HTTP services. The benefits bounded delay-time to carefully opt between the
of delay-bounded routing are follows; (1) the Data Muling and Multihop Forwarding strate-
wireless media resource can be reserved for other gies to minimize communication overhead while
real time service and emergency applications, satisfying with the delay constraints imposed by
and (2) packet delivery does not be affected by the application.
the network density. Packets can be delivered by D-Greedy algorithm adopts only local traffic
carry-and-forward method or multi-hop forward- information to make routing decisions. D-Greedy
ing depended on the constrained delay time. If algorithm chooses the shortest path to destined AP
the current delay time is less than the constrained form the map information, and then allocates the
delay time, carry-and-forward method is used. constrained delay-time to each street within the
Otherwise, multi-hop forwarding method is used shortest path according to the length of street. If
to quickly transmit packets. packets can be delivered under the constrained
Skordylis et al. (2008) proposed a delay- delay-time in a street, Data Muling strategy is
bounded routing protocol in VANETs, which utilized. Packets are carried by a vehicle and
provides a routing scheme that satisfy user-defined forwarded at the vehicle’s speed to destined AP.
delay requirements while at the same time main- Otherwise, Multihop Forwarding strategy is ap-
taining a low level of channel utilization. The plied if packets cannot be delivered within the
delay-bounded routing protocol (Skordylis et al., constrained delay-time. Packets are delivered
216
by multi-hop forwarding. To determine which intersection. The cost of Data Muling strategy is
strategy is adopted, the bounded delay time Del 1, so CDM = 1. The CMH represents the number of
is evaluated as follows, message transmissions in a street by Multihop
Forwarding strategy. The R denotes the maximum
(r - d ) transmission range of a vehicle. The q denotes the
defer_tim e(R ) = MaxDeferTime ´
SR
time required for a vehicle to check its neighbor
r
list and identify the best next hop.
To achieve the minimum cost C within the
constrained delay Del, DSA (Delay Scaling Al-
where TTL is the life time of packet, distToInt is
gorithm) (Goel et al., 2001) is applied, in order
the remaining length, and disToAP is the distance
to efficiently compute delay-constrained least
to AP station. The expected delay time at different
cost paths from the vehicle’s location to all ac-
time is evaluated by the equation below,
cess points on a network. The best routing path is
opted by DSA (Goel et al., 2001) with minimum
distToInt channel utilization under the constrained delay-
DelDM =
u time. Figure 11 shows that Data Muling strategy
A vehicle transmits packets to AP station is applied if the packet can be delivered form VA to
by carry-and-forward method if Del>DelDM. If AP within the constrained delay-time. Otherwise,
Del<DelDM, multi-hop forwarding is applied. the packet is delivered by Multihop Forwarding
D-MinCost algorithm considers the global strategy. When Multihop Forwarding strategy
traffic information in a city to achieve the mini- is applied, the packet is transmitted to a vehicle
mum channel utilization within the constrained which is the closest to AP. Vehicle VC is choose
delay-time. According to the global traffic infor- by VA to relay packets to AP station.
mation, the cost C and delay Del of each street
can be pre-computed. The cost C represents the
number of message transmissions in a street. The mulTicasT and geocasT
delay Del denotes the time required to forward a rouTing proTocol
message in a street. The cost C and delay Del of
a street are evaluated as follows, Multicast and geocast routings are the other impor-
tant routing operations in VANETs. Many VANET
applications require disseminating information to
l
DelDM = ,C = 1, a group of mobile vehicles in a VANET. These
u DM
applications include distributed games, replicated
1 file systems, teleconferencing, etc. The multicast in
C MH = , DelMH + C MH ´ q , VANETs is defined by delivering multicast packets
R
from a mobile vehicle to all multicast-member
vehicles. The geocast in VANETs is defined by
where DelDM denotes the time required to carry- delivering geocast packets from a source vehicle
and-forward a message through a street by Data to vehicles located in a specific geographic re-
Muling strategy, and l denotes the length of a street gion. One of the challenges is how to develop
and u denotes the average vehicle velocity in this the efficient multicast and geocast protocol over
street. When the Data Muling strategy is applied, VANETs with the highly changeable topology and
the packet is carried by a vehicle to forward, the temporary network fragmentation to guarantee that
vehicle only delivers the packet once at next
217
Figure 11. Data muling and multihop forwarding strategies
vehicles can be successfully received the multicast the multicast and geocast protocols. According to
and geocast packets in VANETs. the property of geographic region, existing results
Usually, the temporary network fragmentation can be classified into multicast, geocast protocol,
problem affects the performance of multicast and and spatiotemporary multicast/geocast routing
geocast operations. When a vehicle moves under a protocols as shown in Figure 12. This section
highly speed, the velocity variation between each reviews the existing results for VANETs.
pair of vehicles is large. The distance between
each pair of vehicles quickly changes. This eas- distributed robust geocast
ily offers the temporary network fragmentation multicast routing protocol
problem. When a vehicle easily moves out of the
communication range of the sender and no neigh- The goal of distributed robust geocast multicast
boring vehicle is able to forward packets, then routing protocol is to deliver packets to vehicles
this vehicle fails to receive multicast and geocast located in a specific geographic region. A vehicle
packets. This condition is to lose the connectivity. should receive packets or drop only depended on
Some results (Joshi et al., 2007; Bachir et al., 2003; its current location. If a vehicle is located in the
Chen et al., 2009) have recently investigated the specific geographic region, this vehicle receives
multicast and geocast protocols in a VANET. All packets. Otherwise this vehicle drops packets.
of them mainly consider overcoming the tempo- Packet reception in the geocast multicast routing
rary network fragmentation problem. It is very protocols mainly considers whether vehicles are
important to consider the factors of temporary located in the specific geographic region. Joshi
network fragmentation problem when designing et al. (2007) had proposed a distributed robust
Figure 12.Existing multicast and geocast routing protocols
218
geocast protocol for inter-vehicle communica- overcame with the assistances of vehicles of ZOF,
tion. The zone of relevance (ZOR) is defined in for instance, VG and VF.
(Joshi et al., 2007) as a geographic region which
vehicles in this region should receive the geocast multicast protocol in ad hoc
messages. Joshi et al. pointed out that a stable route networks inter-Vehicle geocast
is difficult to setup in a VANET due to limited
lifetime of a connectivity and frequently network Bachir et al. (2003) proposed a multicast protocol
fragmentation. To enhance the reliability of receiv- in ad hoc networks inter-vehicle geocast, called
ing geocast messages under frequently change- IVG protocol (Bachir et al., 2003). The IVG proto-
able topology, the zone of forwarding (ZOF) is col is used to inform all the vehicles in a highway
defined in (Joshi et al., 2007) as the geographic if any danger is occurred; such as an accident. The
region which vehicles in this region should for- risk area is determined in terms of driving direction
ward the geocast messages to other vehicles in the and positioning of vehicles. Vehicles located in
ZOR. Note that, ZOF usually surrounds ZOR to the risk area form a multicast group. The multicast
ensure the geocast messages can be delivered to group is defined temporarily and dynamically
vehicles inside ZOR. A periodic retransmission by the location, speed, and driving direction of
mechanism is proposed in (Joshi et al., 2007) to vehicles. IVG protocol uses periodic broadcasts
overcome the network fragmentation. Besides, a to overcome temporary network fragmentation
distance-based backoff algorithm is used to reduce for delivering messages to multicast members.
the number of hops and redundant broadcasts. The re-broadcast period is calculated based on the
The random backoff time is determined by the maximum vehicle speed. Besides, IVG protocol
distance to sender vehicle. The longer distance reduces the hops of delivering message by using
is, the smaller backoff time will be, the hops the deferring time, such that
can be reduced. Once a vehicle broadcasts, the
neighboring vehicles may not broadcast, redundant (r - d )
defer_tim e(R ) = MaxDeferTime ´
SR
broadcasts can be reduced. Figure 13 shows that
r
the temporary network fragmentation problem is
Figure 13. DRG multicast routing protocol
219
where r is the transmission range and dSR is the routing protocol should take the time factor into
distance between sender vehicle VS and receiving account. The distinctive feature of this new form
vehicle VR. The farthest vehicle with larger dSR of spatiotemporary multicast and geocast routing
waits for the less time to re-broadcast. Figure 14 protocol is the delivery of information to all nodes
shows an example for the IVG protocol. Vehicle that happen to be in a prescribed region of space
VA encounters car-function-failure problem and at a particular point in time. Chen et al. (2009)
sends this notification to all vehicles in the risk presents a “spatiotemporary multicast”, called a
area. Vehicles VB, VC, and VD form a multicast group “mobicast”, protocol for supporting applications
since they are located in the risk area. Vehicle VC which require spatiotemporary coordination in
is the next hop of VA since the VC is farther from VANETs. The spatiotemporary character of a
VA than VB. After vehicle VC sending out packets, mobicast is to forward a mobicast message to
vehicle VB not forwards packet. vehicles located in some geographic zone at time
t, where the geographic zone is denoted as zone
spatiotemporary multicast/ of relevance (ZORt). Vehicles located in ZORt at
geocast routing protocol the time t should receive the mobicast message.
Many interesting and useful applications on
The spatiotemporary multicast/geocast routing VANETs can be supported by mobicast routing
protocol is a new and very interest routing problem. protocol, such as emergency event (Park et al.,
Unlike ordinary multicast and geocast routing pro- 2006), online game (Palazzi et al., 2007), and
tocols, the spatiotemporary multicast and geocast video advertisement (Yoon et al., 2008). Mobicast
routing protocol can be effectively used for those
VANETs applications for emergency warning,
Figure 14. IVG multicast protocol online game invitation, and video advertisement,
as illustrated in Figure 15. Figure 15 (a) shows
vehicle Ve has the control failure problem for a short
period of time, and warning messages are sent to
all nearby vehicles to avoid accident. By mobicast
routing protocol, warning messages can be sent
to vehicles in the warning area to avoid accident.
Figure 15 (b) shows an online game application,
vehicle Ve can invite those nearby vehicles to be
game-playing members for a longer period of
Figure 15. Applications for mobicast routing
220
time. Our mobicast routing protocol can make Ve and they are indicated to receive the mobicast
sure that the game information can disseminate message by ZORt+1. Observed that although V4. is
to all members in ZORt. out of transmission range of Ve, V4 can receive the
To ensure the mobicast message can be sent mobicast message by relaying from V1. At time
to all vehicles in ZORt, vehicles located in ZORt t + 2, V1 moves away from V4 and V2 moves out
at the time t must keep the connectivity to main- of transmission range of Ve; thus V2 and V4 can
tain the real-time data communication between not receive the mobicast message. The temporal
all vehicles in ZORt. The connectivity is kept of network fragmentation problem occurred on V2
all vehicles in ZORt through the vehicular ad hoc and V4. ZORt is introduced below to solve this
networks. The connectivity of ZORt is lost if any problem.
vehicle in ZORt suddenly accelerates or decelerates To overcome the temporal network fragmentation
its velocity. The temporal network fragmentation problem, ZORt is used to disseminate the mobicast
problem is occurred such that vehicle in ZORt can- message to all vehicles located in the ZORt. ZORt is
not successfully receive the mobicast messages. a geographic zone greater than ZORt to involve more
To solve the problem, Chen et al. proposed a new vehicles to forward the mobicast message. Vehicles
mobicast protocol to successfully disseminate in ZORt should forward the mobicast message to
mobicast messages to all vehicles in ZORt via a another vehicles located in ZORt. ZORt indicates
special geographic zone, called as zone of for- which vehicle should forward the mobicast
warding (ZORt). The proposed mobicast routing message to other vehicles located in the ZORt.
protocol can dynamically estimate the accurate All vehicles in ZORt must forward the received
ZORt. to successfully disseminate mobicast mes- mobicast message; even those vehicles are not
sages to all vehicles in ZORt. As shown in Figure located in ZORt. Figure 17(a) shows V2 and V4 can
16, gives a continuous-time example of ZORi, not receive the mobicast message due to the tem-
where i = t…t + 2, with the temporal network poral network fragmentation problem. Example
fragmentation problem. The transmission range of ZORt is illustrated in Figure 17(b), V5 and V6
of each vehicle is assumed to r. Initially, Ve detects are located in ZORt and have the responsibility
an emergency event at time t to form a ZORt. V1 of forwarding the mobicast message to V2 and V4,
and V3 are involved by this event and indicated respectively. Normally, the size of ZORt may be
to receive the mobicast message by ZORt. V1 and larger or smaller than the optimal size of ZORt . If
V3 are located within the transmission range of Ve; the size of ZORt is larger than the optimal size of
thus Ve directly disseminates the mobicast message ZORt, some irrelevant vehicles are asked to use-
to V1 and V3. At time t + 1, V2 and V4 approach lessly forward the mobicast message. If the size
Figure 16. Temporal network fragmentation problem
221
of ZORt is smaller than the optimal size of ZORt, V

Vi+1 is located in Zt +1
i
. Therefore, multiple zones
the temporal network fragmentation problem is of approaching are initiated to forward the mobi-
incompletely overcome. Observe that, the size of cast message, such that ZORt is finally formed by
ZORt is difficult to predict and determined under all initiated zones of approaching. Therefore, we
the high speed environment, such that it easily V
have ZORt = ZORt∪ Zt i ∪… ∪ Zt i ∪ Zt +1
V Vi
∪… ∪
wastes the network resources. Chen et al. proposed Vn
Zt . Observe that, ZORt is the partial ZORt since
an efficient scheme to estimate the size of ZORt is
the mobicast message should be transmitted to all
near to the optimal size of ZORt. Therefore, zone
V V
vehicles located in ZORt. Figure 18 demonstrates
of approaching ZOAt i or Zt i is proposed herein an example of the proposed mobicast routing
to accurately predict the ZORt. protocol. At time t, V1, V2, V3, and V4 are located
V V
ZOAt i or Zt i is an elliptic zone of approach- in ZORt and receive the mobicast message from
ing to forward the mobicast message more closed vehicle Ve. At time t +1, V2 and V4 can not directly
V
to a destined vehicle and Zt i is initiated by vehicle receive the mobicast messages due to temporal
V
Vi at time t. Any vehicle in the Zt i has the respon- network fragmentation problem. At time t +1, Ve,
Ve V5 V1
sibility of forwarding the mobicast message sent V5, and V1 initiate Zt +1 , Zt +1 , and Zt +1 to forward
V
from vehicle Ve. Zt i bounds the mobicast message the mobicast messages to vehicles V4 and V2. In
Ve V5 V1
V
propagation, vehicles in the Zt i can only forward this case, ZORt+1 = ZORt+1∪ Zt +1 ∪ Zt +1 ∪ Zt +1 .
the mobicast message to other vehicles located in
V
the Zt i . If a vehicle cannot successfully forward
the mobicast message to any neighbor vehicle broadcasT rouTing proTocol
V
in the Zt i which is more closed to the destined
Broadcast is the last important operation for a
vehicle, a new approaching zone is initiated. The
Vi vehicle to disseminate a broadcast message to
growing of a new zone of approaching Zt +1 from all the others in a VANET. To overcome the tem-
V
an existing zone of approaching Zt i is explained as porary network fragmentation problem to avoid
follows. Given two connected approaching zones some vehicles cannot be successfully received the
V
Zt i and ZtV+1
i
V
, if Vi+1 in Zt i cannot successfully broadcast message. This section describes existing
forward the mobicast message to any neighbor broadcast routing protocols in VANETs, as listed
vehicle closed to the destined vehicle, then Vi+1 in Figure 19, as follows.
Vi
initiates a new zone of approaching Zt +1 , where
Figure 17. Function of ZORt and ZOFt
222
Figure 18. Zone of approaching ZORt
Figure 19. Existing broadcast routing protocols
on The broadcast storm problem persistence schemes. In the weighted p-persistence

in ad hoc wireless networks scheme, if vehicle Vj receives a packet form vehicle
Vi, vehicle Vj first checks whether the packet has
The broadcast storm problem has been heavily been received. If vehicle Vj receives this packet
discussed in MANETs (Tseng et al., 2002). Tonguz at the first time, then vehicle Vj has probability
et al. (2006) further demonstrates the broadcast pij to re-broadcast the packet. Otherwise, vehicle
storm problem occurred in VANETs. The broad- Dij
cast storm problem causes serious packet collision Vj drops this packet, under pij = , where Dij
R
and packet loss since too many vehicles simulta- is the distance between vehicle Vi and Vj, R is
neously broadcast messages. A shortest routing the transmission range. Neighbors of vehicle Vi
path can be discovered by broadcasting the RREQ change pij to 1 to ensure that the message must
(route request) message. The broadcast routing be broadcasted if they have no received the re-
strategies developed in (Tonguz et al., 2006) are broadcast message after waiting a random time.
1-persistence and p-persistence. However, these In the slotted 1-persistence scheme, If vehicle Vj
two strategies easily cause the broadcast storm firstly receives this packet from vehicle Vi, then
problem, especially in a high density network. vehicle Vj waits for TS time slots, vehicle Vj has
Tonguz et al. (2006) thus proposed three distrib- ij
probability 1 to re-broadcast the packet, where
uted broadcast suppression techniques, weighted
TS = Sij×τ.
p-persistence, slotted 1-persistence, and slotted p- ij
223
ïìïé æ D öù dV-casT: broadcasting in VaneT

ïïêêN ççç1 - ij ÷÷÷úú , if Dij
Sij = íê s çè R ÷÷øú ,
ïïê ú Tonguz et al. (2007) proposed DV-CAST for a
ïï 0, if Dij > R multi-hop broadcast routing protocol in VANETs.
ïî
Tonguz et al. (2007) indicates three traffic sce-
where τ is the propagation time for one hop narios for a vehicular broadcasting; (1) dense
transmission and Ns is the default number of traffic scenario, (2) sparse traffic scenario, and
time-slot. (3) regular traffic scenario. Tonguz et al. (2007)
The slotted p-persistence scheme combines the integrates previously proposed routing solution
weighted p-persistence and slotted 1-persistence in (Tonguz et al., 2006) to develop DV-CAST
schemes. If vehicle Vj firstly receives the packet which is suitable for both of dense and sparse
from Vi, then vehicle Vj waits for TS time-slots. traffic scenarios. The overhead in dense traffic
ij
Vehicle Vj has probability pij to re-broadcast the scenario can be reduced. DV-CAST utilizes the
packet. Simulation results show that these three information form one-hop neighbors to make the
schemes can achieve up to 90% reduction of routing decision. Each vehicle records the states of
packet loss rate. Figure 20 gives the example of neighboring vehicles all the time. Three parameters
these three broadcast schemes. In Figure 20(a), should be recorded; (1) DFlag (Destination Flag),
the weighted p-persistence scheme is used. Figure (2) MDC (Message Direction Connectivity), and
20(b) illustrates the slotted 1-persistence scheme. (3) ODC (Opposite Direction Connectivity). The
Figure 20(c) shows the slotted p-persistence DFlag indicates whether the direction of the cur-
scheme. rent vehicle is the same as the source vehicle. The
MDC describes whether a vehicle exists behind
Figure 20. Broadcast suppression techniques (a) weighted p-persistence scheme (b) slotted 1-persistence
scheme (c) slotted p-persistence scheme
224
Figure 21. Well-connected neighborhood
the current vehicle to assist broadcast message broadcast methods for

forwarding. The ODC indicates whether a vehicle inter-Vehicle communications
exists in the opposite direction. Vehicles can make system
decision based on these three parameters. For
example, if a vehicle Vi receives a new broadcast Fukuhara et al. (2005) proposed broadcast meth-
message, vehicle Vi firstly checks MDC. If there ods for inter-vehicle communications system to
are vehicles behind vehicle Vi(MDC=1), vehicle Vi provide emergency information dissemination in
uses broadcast suppression schemes proposed in VANETs. The purpose of emergency information
(Tonguz et al., 2006) to forwarding the broadcast is to announce an urgent event by broadcasting for
message. Otherwise (MDC=0), vehicle Vi checks surrounding vehicles. According to the purposes
ODC. If there are vehicles in the opposite direction of emergency information, the proposed broadcast
(ODC=1, MDC=0), vehicle Vi forwards the broad- methods in (Fukuhara et al., 2005) are divided
cast message to vehicles in the opposite direction. into two categories, emergency-vehicle-approach
After vehicle Vi broadcasting message, vehicle information and traffic accident information.
Vi checks DFlag. If the direction of vehicle Vi is Emergency-vehicle-approach information is used
different from source vehicle (DFlag=0), vehicle to announce the urgent event to those vehicles
Vi overhears for a period of time to ensure that in front of the current vehicle, so the emergency
the message is successfully broadcasted. Figure information is only disseminated ahead. Traffic
21 shows that the broadcast message is initiated accident information is used to announce the
by VS and it is forwarded from group 1 to group urgent event to those vehicles behind the cur-
2. Although groups 1, 2, and 3 are dense group, rent vehicle; the emergency information is only
groups 1 and 2 encounter the temporary network disseminated behind. By limiting the broadcast
fragmentation problem. Group 1 cannot directly direction, the proposed broadcast methods (Fu-
forward packets to group 2. In this case, vehicle kuhara et al., 2005) can provide broadcasts to a
VS can forward packets to group 3 which is in the particular area and avoid mistakenly notifying
reverse direction, then vehicle VS forwards packets other areas where the information is not needed.
to group 2. Observe that, the temporary network To decide emergency information should be re-
fragmentation problem is also considered in the ceived, forwarded, or dropped, three conditions
design of broadcasting. below are examined:
   
(X C )
- X E + aV F ×V F > 0 , (1)
  
X C - X E + aV F < Rnotification , (2)
225
Figure 22. The broadcast area for emergency information
  
X C - X E + aV F < Rrelay , sible. The delay-bound routing protocol utilizes
(3)
the carry-and-forward technique to minimize the
channel utilization within a constrained delay time.
This chapter also surveys important multicast and
where XC is the location vector of the current ve-
geocast protocols for VANETs. The multicast in
hicle, XE is the location vector of the emergency
VANETs is defined by delivering multicast packets
vehicle, VF is the forward direction vector of the
from a mobile vehicle to all multicast-member
emergency information, Rnotification is the available
vehicles. The geocast in VANETs is defined by
notification range, and Rrelay is the available relay
delivering geocast packets from a source vehicle
range.
to vehicles located in a specific geographic region.
If conditions (1) and (2) are true, the current
A mobicast routing protocol in VANETs is also
vehicle receives the emergency information. If
described. Finally, broadcast protocols in VANETs
conditions (1) and (3) are true, the current vehicle
are also introduced. We predict the tendency of
re-broadcasts the emergency information. Figure
the design of routing protocols for VANETs must
22 shows that vehicle VA broadcasts the emergency
be the low communication overhead, the low time
message to the restricted direction. Vehicle VD does
cost, and high adjustability for the city, highway,
nothing. Vehicle VD is located in the relay range, it
and rural environments.
re-broadcasts the emergency information. Vehicle
VD is located in notification range but not in relay
range, VC just receives the emergency information
reFerences
and not to re-broadcast.
Amouris, K. (2005). Position-Based Broadcast
TDMA Scheduling for Mobile Ad-Hoc Networks
conclusion (MANETs) with Advantaged Nodes. IEEE Mili-
tary Communications Conference, 1, 252–257.
Unicast, multicast, and broadcast routing op-
erations are key issues in the network layer for ASTM International E2213-03 (2003). Stan-
VANETs. This chapter surveys existing unicast, dard Specification for Telecommunications and
multicast, and broadcast protocols for VANETs. Information Exchange Between Roadside and
The unicast routing protocols are split into Vehicle Systems 5 GHz Band Dedicated Short
min-delay and delay-bound approaches. The Range Communications (DSRC) Medium Ac-
min-delay unicast routing protocols construct cess Control (MAC) and Physical Layer (PHY)
a minimum-delay routing path as soon as pos- Specifications.
226
Bachir, A., & Benslimane, A. (2003). A Multi- Hughes, L., & Maghsoudlou, A. (2006). An Ef-
cast Protocol in Ad hoc Networks Inter-Vehicle ficient Coverage-Based Flooding Scheme for
Geocast. IEEE Semiannual Vehicular Technology Geocasting in Mobile Ad Hoc Networks. Inter-
Conference, 4, 2456-2460. national Conference on Advanced Information
Networking and Applications, 1, 6.
Blaževi’c, L., Giordano, S., & LeBoudec, J.
Y. (2002). Self Organized Terminode Routing. Joshi, H. P., Sichitiu, M., & Kihl, M. (2007).
Cluster Computing Journal, 5(2), 205–208. Distributed Robust Geocast Multicast Routing for
doi:10.1023/A:1013998030317 Inter-Vehicle Communication. WEIRD Workshop
on WiMax, Wireless and Mobility.
Blum, J., Eskandarian, A., & Hoffman, L. (2004).
Challenges of inter-vehicle ad hoc networks. Karp, B. (2001). Challenges in Geographic Rout-
IEEE Transactions on Intelligent Transpor- ing: Sparse Networks, Obstacles, and Traffic
tation Systems, 5(4), 347–351. doi:10.1109/ Provisioning. In DIMACS Workshop on Pervasive
TITS.2004.838218 Networking.
Briesemeister, L., & Hommel, G. (2000). Over- Karp, B., & Kung, H. T. (2000). GPSR: Greedy
coming Fragmentation in Mobile Ad hoc Net- Perimeter Stateless Routing for Wireless Net-
works. Journal of Communications and Networks, works. ACM International Conference on Mobile
2(3), 182–187. Computing and Networking, (pp. 243-254).
Chen, Y. S., Lin, Y. W., & Lee, S. L. (2009). A Lochert, C., Mauve, M., Füßler, H., & Harten-
Mobicast Routing Protocol for Vehicular Ad Hoc stein, H. (2005). Geographic routing in city
Networks. Mobile Networks and Applications scenarios. ACM SIGMOBILE Mobile Com-
(MONET).New York: ACM/Springer. puting and Communications, 9(1), 69–72.
doi:10.1145/1055959.1055970
Chen, Y. S., Lin, Y. W., & Pan, C. Y. (2009). A
Diagonal-Intersection-Based Routing Protocol Manoharan, R., & Thambidurai, S. L. P. P. (2008).
for Urban Vehicular Ad Hoc Networks. Telecom- Energy Efficient Robust On-Demand Multicast
munication System, (under revision). Routing Protocol for MANETs. [IJAHUC].
International Journal of Ad Hoc and Ubiq-
Fukuhara, T., Warabino, T., Ohseki, T., Saito, K.,
uitous Computing, 3(2), 90–98. doi:10.1504/
Sugiyama, K., Nishida, T., & Eguchi, K. (2005).
IJAHUC.2008.017002
Broadcast Methods for Inter-vehicle Communica-
tions System. IEEE Wireless Communications and Naumov, V., Baumann, R., & Gross, T. (2006).
Networking Conference, 4, 2252-2257. An Evaluation Of Inter-Vehicle Ad Hoc Networks
Based On Realistic Vehicular Traces. In ACM
Gerten, G. (2005). Protecting the Global Posi-
International Symposium on Mobile Ad Hoc
tioning System. IEEE Aerospace and Electronic
Networking and Computing, (pp. 108-119).
Systems Magazine, 20(11), 3–8. doi:10.1109/
MAES.2005.1576067 Naumov, V., & Gross, T. (2007). Connectivity-
Aware Routing (CAR) in Vehicular Ad Hoc
Goel, A., Ramakrishnan, K. G., Kataria, D., &
Networks. IEEE International Conference on
Logothetis, D. (2001). Efficient Computation of
Computer Communications (INFOCOM’07),
Delay-Sensitive Routes from One Source to All
(pp. 1919-1927).
Destinations. IEEE Conference on Computer
Communications (INFOCOM), (pp. 854–858).
227
Palazzi, C. E., Roccetti, M., Pau, G., & Gerla, Tonguz, O., Wisitpongphan, N., Bai, F., Mudalige,
M. (2007). Online Games on Wheels: Fast Game P., & Sadekar, V. (2007). Broadcasting in VANET.
Event Delivery in Vehicular Ad-hoc Networks. In Mobile Networking for Vehicular Environments,
IEEE Intelligent Transportation Systems Society, (pp. 7-12).
(pp. 42–49).
Tonguz, O. K., Wisitpongphan, N., Parikh, J. S.,
Park, J. S., Lee, U., Oh, S. Y., Gerla, M., & Lun, Bai, F., Mudalige, P., & Sadekar, V. K. (2006).
D. (2006). Emergency Related Video Stream- On the Broadcast Storm Problem in Ad hoc
ing in VANETs using Network Coding. In ACM Wireless Networks. International Conference
International Workshop on Vehicular Ad Hoc on Broadband Communications, Networks and
Networks (VANET), (pp. 102-103). Systems, (pp. 1-11).
Safa, H., Artail, H., & Shibli, R. (2009). An Tseng, Y. C., Ni, S. Y., Chen, Y. S., & Sheu,
Interoperability Model for Supporting Reliabil- J. P. (2002). The Broadcast Storm Problem
ity and Power-Efficient Routing in MANETs. in a Mobile Ad Hoc Network. [WINET].
[IJAHUC]. International Journal of Ad Hoc and ACM Wireless Networks, 8(2), 153–167.
Ubiquitous Computing, 4(2), 71–83. doi:10.1504/ doi:10.1023/A:1013763825347
IJAHUC.2009.023898
Wan, S., Tang, J., & Wolff, R. S. (2008). Reliable
Saha, A. K., & Johnson, D. B. (2004). Modeling Routing for Roadside to Vehicle Communications
mobility for vehicular ad hoc networks. ACM in Rural Areas. In IEEE International Conference
International Workshop on Vehicular Ad Hoc on Communications, (pp. 3017-3021).
Networks (VANET), (pp. 91-92).
Xu, Q., Mak, T., Ko, J., & Sengupta, R. (2004).
Sawamura, T., Tanaka, K., Atajanov, M., Matsu- Vehicle-to-Vehicle Safety Messaging in DSRC.
moto, N., & Yoshida, N. (2008). Adaptive Router In ACM International Workshop on Vehicular Ad
Promotion and Group Forming in Ad-Hoc Net- Hoc Networks (VANET), (pp. 19-28).
works. [IJAHUC]. International Journal of Ad
Yoon, H., Kim, J., Tan, F., & Hsieh, R. (2008).
Hoc and Ubiquitous Computing, 3(4), 217–223.
On-demand Video Streaming in Mobile Opportu-
doi:10.1504/IJAHUC.2008.018907
nistic Networks. In IEEE International Conference
Skordylis, A., & Trigoni, N. (2008). Delay- on Pervasive Computing and Communications
Bounded Routing in Vehicular Ad-Hoc Networks. (PerCom), (pp. 80–89).
ACM international symposium on Mobile ad hoc
Yousefi, S., Mousavi, M. S., & Fathy, M. (2006).
networking and computing, (pp. 341-350).
Vehicular Ad Hoc Networks (VANETs): Chal-
Sun, W., Yamaguchi, H., Yukimasa, K., & lenges and Perspectives. International conference
Kusumoto, S. (2006). GVGrid: A QoS Routing on ITS Telecommunications, (pp. 761-766).
Protocol for Vehicular Ad Hoc Networks. IEEE
Zhao, J., & Cao, G. (2006). VADD: Vehicle-
International Workshop on Quality of Service,
Assisted Data Delivery in Vehicular Ad Hoc
(pp. 130-139).
Networks. IEEE Computer Communications,
Taleb, T., Sakhaee, E., Jamalipour, A., Hashimoto, (pp. 1-12).
K., Kato, N., & Nemoto, Y. (2007). A Stable Rout-
ing Protocol to Support ITS Services in VANET
Networks. IEEE Transactions on Vehicular
Technology, 56(6), 3337–3347. doi:10.1109/
TVT.2007.906873
228
229
Chapter 14
Applications in Vehicular
Ad Hoc Networks
Tzung-Shi Chen
National University of Tainan, Tainan, Taiwan, R.O.C.
Hua-Wen Tsai
Yi-Shiang Chang
absTracT
The various sensors and wireless communication devices have been extensively applied to daily life due
to the advancements of microelectronics mechanism and wireless technologies. Recently, vehicular com-
munication systems and applications become more and more important to people in daily life. Vehicular
communication systems that can transmit and receive information to and from individual vehicles have
the potential to significantly increase the safety of vehicular transportation, improve traffic flow on con-
gested roads, and decrease the number of people of deaths and injuries in vehicular collisions effectively.
This system relies on direct communication between vehicles to satisfy the communication needs of a
large class of applications, such as collision avoidance, passing assistance, platooning. In addition,
vehicular communication systems can be supplemented by roadside infrastructure to access Internet and
other applications. This system forms a special case of mobile ad hoc networks called Vehicle Ad Hoc
Networks (VANETs). They can be formed between vehicles with vehicle to vehicle (V2V) communica-
tion or between vehicles and an infrastructure with vehicle to infrastructure (V2I) communication. The
applications and characteristics of VANETs are introduced and presented in this Chapter.
inTroducTion electronics mechanism and wireless technologies.

Vehicular transportation is one of the main modes
Recently, the various sensors and wireless com- of transportation for millions of U.S. citizens and
munication devices have been extensively applied hundreds of millions around the world in spite of
to daily life due to the advancements of micro- increasing oil prices and environmental concerns.
However, 38,252 Americans were killed in 2003
with 2,697,000 seriously injured. Furthermore,
DOI: 10.4018/978-1-60566-840-6.ch014
Applications in Vehicular Ad Hoc Networks
each day the average American spends 2.5 hours reporting data to a central database, or cellular
in his/her vehicle, a significant percentage of wireless communication between vehicles and
this time in traffic jams and at stop lights. The a monitoring center. The advantage of VANET
statistics are similar in many other parts of the is not needed to deploy too many expensive
world (Sichitiu et al., 2008). On an average day infrastructures to be installed on every road and
in the United States, vehicular collisions kill intersections. VANETs are emerging as the pre-
116 and injure 7900. More health care dollars ferred network design for intelligent transportation
are consumed in the United States treating crash systems. VANETs are based on short-range wire-
victims than any other cause of illness or injury; less communication (e.g., IEEE 802.11) between
the situation in the European Union is similar, vehicles. However, the disadvantage of VANETs is
with over 100 deaths and 4600 injuries daily, and in supporting complex network protocols. The ex-
the annual cost of € 160 billion. Governments and isting researches focus on medium access control
automotive companies are responding by making (MAC) and routing issues, in particular pointing
the reduction of vehicular fatalities a top priority to the mismatch between the need of inter-vehicle
(Robinson et al., 2006). communication applications for group communi-
Vehicular communication systems that can cations and the services offered by mobile ad hoc
transmit and receive information to and from network (MANET) routing protocols.
individual vehicles have the potential to signifi- Several applications enabled by vehicular
cantly increase the safety of vehicular transpor- communication systems are classified. The most
tation, improve traffic flow on congested roads, commonly considered applications are related
and decrease the number of people of deaths and to public safety and traffic coordination. Traffic
injuries in vehicular collisions effectively. This management applications, traveler information
system relies on direct communication between support, and various comfort applications have
vehicles to satisfy the communication needs the potential to make travelling more efficient,
of a large class of applications (e.g., collision convenient and pleasant. For each type of appli-
avoidance, passing assistance, platooning, and cation, we consider its addressing and real-time
so on). Vehicular communication systems can be requirements and the type of vehicular commu-
supplemented by roadside infrastructure to access nication necessary for its implementation.
Internet and other applications. This system forms Generally speaking, vehicular communication
a special case of mobile ad hoc networks called systems are classified into three major types,
Vehicle Ad Hoc Networks (VANETs). They can be inter-vehicle communication (IVC), roadside-
formed between vehicles with vehicle to vehicle to-vehicle communication systems (RVC), and
(V2V) communication or between vehicles and hybrid vehicular communication (HVC).
an infrastructure with vehicle to infrastructure IVC systems are completely infrastructure-free
(V2I) communication. and only need some in-vehicle equipments. De-
VANETs can increase security, efficiency and pending on whether the information is retransmit-
comfortable trip in collecting road traffic. Such ted at intermediate hops or not, we can further dis-
systems can enable a wide range of applications, tinguish between single-hop and multi-hop IVCs,
such as collision avoidance, emergency message SIVCs and MIVCs, respectively. SIVC systems
dissemination, dynamic route scheduling, route are useful for applications requiring short-range
navigation, and real-time traffic condition moni- communications, e.g., lane merging, automatic
toring. Traditional vehicular networks for report- cruise control. MIVC systems are more complex
ing accidents or traffic conditions rely on certain than SIVCs but can also support applications that
infrastructure, such as road-side traffic sensors require long-range communications, e.g., traffic
230
monitoring. An MIVC system requires a network Research Institute (JSK) are a major step towards
layer capable of multi-hop routing. the realization of Intelligent Transport Services
In RVC systems, all communications take place (ITS) (Taleb et al., 2007).
between roadside infrastructure and vehicles. Dedicated short range communications
Depending on the application, two different types (DSRC) standard at 5.9 GHz band is projected to
of infrastructure can be classified, sparse RVC support low-latency wireless data communications
(SRVC) and ubiquitous RVC (URVC) systems. between vehicles and from vehicles to roadside
SRVC systems are capable of supporting com- units. The DSRC specification is meant to be an
munication services at hot spots, for examples, extension of the IEEE 802.11 technology into the
a busy intersection scheduling its traffic light, outdoor high-speed vehicle environment. In fact,
a restaurant advertising its existence and prices, the physical layer (PHY) of DSRC is adapted from
and parking availability at an airport. An SRVC IEEE 802.11a PHY based on orthogonal frequency
system can be deployed gradually, thus not requir- division multiplex (OFDM) technology. More-
ing substantial investments before any available over, the multiple access control (MAC) layer of
benefits. A URVC system is the holy grail of DSRC is very similar to the IEEE 802.11 MAC
vehicular communication in which providing based on the carrier sense multiple access with
all roads with high-speed communication would collision avoidance (CSMA/CA) protocol with
enable applications unavailable with any of the some minor modifications (Ryu et al., 2004).
other systems. Unfortunately, a URVC system may MANETs are wireless multi-hop networks,
require considerable investments for providing decentralized and self-organized without infra-
full coverage of existing roadways, especially in structure. VANETs satisfy all these requirements,
large countries (Sichitiu et al., 2008). and are a special class of MANETs. However, the
HVC systems are proposed for extending followings are several characteristics that differ-
the range of RVC systems. In HVC systems ve- entiate between VANETs and MANETs (Sichitiu
hicles communicate with roadside infrastructure et al., 2008) .
even when they are not in direct wireless range
by using other vehicles as mobile routers. The • Applications: The most common assump-
main advantage is that it requires less roadside tion of MANET applications are identical
infrastructure. However, one disadvantage is that to those enabled by the Internet. In con-
network connectivity may not be guaranteed in trast, VANETs have completely different
scenarios with low vehicle density. applications to increase the safety and
The Federal Communications Commission comfortable trip. VANET applications in-
(FCC) has allocated spectrum for IVC and clude onboard active safety systems to as-
similar applications (e.g., Wireless Access in sist drivers in avoiding collisions and to
Vehicle Environment, WAVE). Governments and coordinate among them at critical points
prominent industrial corporations, such as Toyota, such as intersections and highway entries.
BMW, and Daimler-Chrysler have launched Safety systems may intelligently dissemi-
important projects for IVC communications. nate road information, such as incidents,
Some notable projects, e.g., Advanced Driver real-time traffic congestion, high-speed
Assistance Systems (ADASE2), Crash Avoidance tolling, or surface condition to vehicles
Metrics Partnership (CAMP), Chauffeur in EU, in the vicinity of the subjected sites. This
CarTALK2000, FleetNet, California Partners helps to avoid platoon vehicles and to ac-
for Advanced Transit and Highways (California cordingly improve the roads capacity. With
PATH), and DEMO 2000 by Japan Automobile such active safety systems, the number of
231
car accidents and associated damage are • A central processing unit (CPU) that imple-
expected to be largely reduced. ments the applications and communication
• Receiver: MANET applications require protocols;
point-to-point (unicast) with fixed address- • A wireless transceiver that transmits and
ing. The recipient of a message is another receives data to/from the neighboring ve-
node in the network specified by its IP ad- hicles and roadside units;
dress. However, VANET applications often • A GPS receiver that provides relatively ac-
require dissemination of the messages to curate positioning and time synchroniza-
many nodes (multicast) that satisfy some tion information;
geographical constraints and possibly oth- • Appropriate sensors to measure the various
er criteria (e.g., direction of movement). parameters that have to be measured and
The need for this addressing mode requires eventually transmitted;
a significantly different routing paradigm. • An input/output interface that allows hu-
• Rate of link changes: The nodes in man interaction with the system.
MANETs are assumed to have moderate
mobility. This assumption allows MANET
routing protocols to establish end-to-end applicaTions in VaneTs
paths that are valid for a reasonable amount
of time and only occasionally need repairs. Based on the different types of vehicular com-
In VANETs applications, it is shown that munication systems (e.g., IVC and RVC), the
due to the high degree of mobility of the telematics have five different classifications in
nodes involved, even multi-hop paths that the applications shown in Figure 1 (Sichitiu et
only use nodes moving in the same direc- al., 2008). The applications in VANETs consist of
tion on a highway have a lifetime compara- the public safety application, traffic management
ble to the time needed to discover the path. applications, traffic coordination and traffic as-
Therefore, the communication of VANETs sistance, traveler information support, and comfort
has more challenge than that of MANETs. applications. The classifications of these applica-
• Mobility model: In MANETs, the random tions are not meant to the overall applications in
waypoint (RWP) is the most commonly VANETs. However, the adopted classification can
employed mobility model. However, for represent the greater part of the current general
VANETs, most existing literatures recog- applications in VANETs. The explanations about
nized that RWP would be a very poor ap- the applications will be described as follows in
proximation of real vehicular mobility; in- detail.
stead, detailed vehicular traffic simulators Public safety applications are very important
are used. issue in the Telematics. Safety applications are
• Energy efficiency: While in MANETs a geared primarily toward avoiding accidents and
significant body of literatures is concerned loss of life of the occupants of vehicles. Collision
with power-efficient protocols, VANETs warning systems have been proposed to reduce
enjoys a practically unlimited power the number of vehicle collisions in several sce-
supply. narios.
On highways, frontal collisions with slow
In VANETs, the hardware available in each moving (or stopped) vehicles are one of the most
vehicle is listed below. common types of accidents, often with serious
consequences. A vehicle with its airbags deployed,
232
Figure 1. Telematics applications
or a stopped or rapidly decelerating vehicle can A similar situation may occur with other types
transmit warning signals to all other approaching of vehicles (e.g., trains). In some cases, if a col-
vehicles (shown in Figure 2(a)). Intermediate lision is imminent, the system may be able to
relays may be used to increase the dissemination prepare the vehicles for collision (inflate air bags,
range of the warning beyond the direct transmis- tighten seat belts, etc.). Thus, safety applications
sion range. have obvious real-time constraints, as drivers
At junctions, vehicles running red stop lights have to be notified before the information is no
often result in side crashes. If both vehicles to longer useful. In terms of addressing, the destina-
be involved in the accident are equipped with tions in these applications will not be individual
vehicular communication systems, such accidents vehicles, but rather any relevant vehicle. The zone
can be prevented (shown in Figure 2(b)). of relevance (ZOR) (also known as the target
area) is determined by the particular application.
Figure 2. Safety applications
233
For example, an accident in the right lane of a presence and intended routes (shown in Figure
highway will only affect vehicles approaching the 3(b)).
accident from behind, while at an intersection the Traffic light scheduling can be significantly
vehicles with intersecting trajectories and speeds improved by using an SRVC (Sparse Roadside-
over a threshold are relevant. to-Vehicle Communication) system. Currently,
Traffic management applications are fo- many traffic lights are scheduled either statically
cused on improving traffic flow, thus reducing or only considering limited information (e.g., by
both congestion as well as accidents resulting sensing the presence or absence of a vehicle in front
from congestion, and reducing travel time. The of a traffic light). An SRVC system can provide
traffic management applications include the additional information, such as the length of the
traffic monitoring, traffic light scheduling, and queues at the traffic light as well as the number
emergency vehicles. of vehicles expected to arrive in the near future,
Traffic monitoring can provide high-resolution which can improve the efficiency of schedules.
localized timely information regarding the traffic Applications in this class generally do not
for several miles around the current location of have stringent real-time requirements: the qual-
the vehicle. For this application each vehicle in ity of the information degrades gracefully with
the system will act as a sensor (determining its the increase of delay and packet loss. Similar to
current speed), as a relay (if the information is to the case of safety applications, the destinations
travel for more than the direct transmission range) of the information are any vehicles in the ZOR.
as well as a destination (using information from For traffic monitoring applications the ZOR can
the other vehicles in the system (shown in Figure be several miles around the information source.
3(a))). The information can be used to simply For traffic light scheduling, traffic lights being
inform the driver or, in more complex systems, to approached are appropriate destinations.
reroute, estimate the time to destination, or even Traffic coordination and traffic assistance
control the traffic by using adaptive speed limits, have been the main research topics of many IVC
ramp metering, and so on. projects. Platooning (i.e., forming tight columns
Emergency vehicles may notify the relevant of vehicles closely following each other on high-
vehicles as well as equipped stop lights of their ways) has the potential to radically increase the
capacity of existing highways. High-speed closed
Figure 3. Traffic management applications
234
loop control is of paramount importance for this tion. Penetration directly influences the usability
application (shown in Figure 4(a)). Passing and of these systems.
lane change assistance may reduce or eliminate Traveler information support applications
risks during these maneuvers, since they are provide updated local information, maps, and in
often the source of serious accidents (shown in general messages of relevance limited in space
Figure 4(b)). and/or time. These messages mainly focus on the
Clearly these applications require close-range local information and road warnings.
IVC with tight real-time constraints and can Local information such as local updated maps,
be implemented with either an SIVC (Sparse the location of gas stations, parking areas, and
Roadside-to-Vehicle Communication) or a URVC schedules of local museums can be downloaded
(Ubiquitous Roadside-to-Vehicle Communica- from selected infrastructure places or from other
tion) system. Both systems can offer similar “local” vehicles. Advertisements with, for ex-
real-time guarantees and delays if properly de- ample, gas or hamburger prices may be sent to
signed, although an SIVC system may have a approaching vehicles (shown in Figure 5(a)).
slight advantage as it faces reduced contention Road warnings of many types (e.g., ice, oil,
and direct links. These applications also have ad- or water on the road, low bridges, or bumps)
dressing based on ZOR; for example, immediately may easily be deployed by authorities simply by
behind, in the right lane, or in the reverse direc- dropping a beacon in the relevant area (shown in
Figure 5(b)).
Figure 4. Traffic coordination and traffic assistance
Figure 5. Traveler information support applications
235
Figure 6. Communicate with different targets
Only a few papers consider specific traveler of applications, from email and media streaming
information support applications. All information to Web browsing and voice over IP. Unfortunately,
support applications require an SRVC system. IVC land-based access requires a URVC system that
systems may augment the service provided by the may be prohibitively expensive in the near fu-
SRVC, but cannot replace it. No special real-time ture.
requirements are necessary, and the penetration Finally, there are many other comfort applica-
percentage has no impact on usability. A vehicle tions. Tolls for roads and bridges can be collected
equipped with a vehicle communication system automatically. Many nonstandard systems exist
benefits from the applications independent of the and work well. Parking payments can be made
OBU penetration rate. The addressing is once promptly and conveniently. Repair and main-
again based on relevance rather than individual tenance records can be recorded at the garages
vehicle IDs. performing them. Multimedia files such as DVDs,
The main focus of comfort applications is to music, news, audiobooks, pre-recorded shows can
make travel more pleasant. This class of applica- be uploaded to the car’s entertainment system
tions may be motivated by the desire of passengers while the car is in the garage.
to communicate with either other vehicles or
ground-based destinations such as Internet hosts or
the public service telephone network (PSTN). TraFFic inFormaTion in VaneTs
Targeted vehicular communications allow
localized communications (potentially multi-hop) According to the definition of Hudson Valley
between two vehicles (shown in Figure 6(a)). Transportation Management Center (HVTMC)
Voice, instant messaging, or similar communica- (“Hudson Valley Transportation Management
tions may occur between the occupants of vehicle Center (HVTMC),”), traffic information means
caravans traveling together for long distances, dynamic information concerning about traffic.
or between law enforcement vehicles and their Thus, traffic information in VANETs means the
“victims.” Note that this application does not dynamic information obtained from the on-board
scale to large network sizes. units in the vehicles. Compared to the travel in-
Vehicle to land-based destination communica- formation, like the maps of the streets or the maps
tions (shown in Figure 6(b)) is arguably a very of the bus routes, traffic information in VANETs
useful capability as it may enable an entire array is usually more immediate and local. As follow-
236
Figure 7. Relevance of spatial-temporal speed

ing, we will introduce and discuss some reference
papers about the collections, presentations, and
applications of the data, such as immediate road
traffic information and navigation information
etc.
surface street Traffic estimation
Average velocity is the most commonly used

parameter to show the traffic information, such as
the real-time traffic network systems of cities and
counties in Taiwan (“Real-Time Traffic Network
Systems of Cities and Counties in Taiwan,”), all
provide average velocity of road sections as an
immediate reference. And the literature “Surface
Street Traffic Estimation” (Noble et al., 2007) estimating Velocity Fields
provides a simple way to determine the current on a Freeway from low-
road conditions by using data analysis through the resolution Videos
GPS and the mobile communication systems. The
approach is as follows: First of all, split the road This is an approach using low-resolution video
section into small segments, and drive a car passing to estimate the velocity fields (Rice et al., 2006).
through the same route repeatedly to collect the Nowadays, there are many image processing
road information (record the relationship between techniques used in traffic surveillance (Kastinaki
time and location). Secondly, update each GPS et al., 2003) . However this algorithm is not only
track data to the central server, and use the spatial- used to estimate the speed for a single vehicle,
temporal traffic status plot to describe the traffic nor to obtain the estimated value of the derivative
characteristics, and then obtained the temporal through the intensity of each point, but rather to
velocity- the average speed of the journey and the find out the most matching value by calculating
spatial velocity- the average speed of the section. the “intensity.” It is divided into two parts as
Thirdly, they use Quadrant Clustering to classify intensity profiles and speed estimation. The first
the spatial-temporal traffic status plot to predict part, intensity profiles, is to choose four related
the present traffic condition by the classification fixed objects from the video (Figure 14.8) between
of the data. Then present the traffic condition as the time interval [t – τ t + τ,], then mark them
good, normal (spatially good but temporally bad, or out with the frame, next find the best matching
temporally good but spatially bad) and poor. And pattern in each frame, strengthen its intensity
then transmit the information acquired through the and use the centric coordinates to compute the
internet. (Figure 7) Finally, transmit the packages projection matrix. Repeat the process. After some
through the internet. This method presents the road time, the intensity flow can be depicted so as to
traffic situation with the effective spatial-temporal know the traffic conditions. In the second part,
plot, taking the waiting time of traffic lights and speed estimation, the basic idea is that the pattern
intersections into account. of the intensity can present the speed of vehicles
237
Figure 8. Filtered image and the corresponding intensity profile
in the region. So the same pattern will be found flow will show up at the same traffic flow. Thus,
at the next point in a short period. The processes this paper provides a algorithm which only uses
are as follows: Fix the chosen frame, find out the a single vehicle to estimate local density, and
intensify relationship d() between time instant t - determine whether it is the case of congestion,
τk and t + τk, where located x - d and x + d. Then then apply the result on the transmission range
take Bayesian point to estimate the velocity and dynamically to meet the need of maintaining the
get the global vehicles’ velocity. high-connectivity of the network. Based on the
two-fluid model (Robert et al., 1979), compute
local density estimation fraction (f) = the stopping time of test vehicle (Ts)
and dynamic Transmission- divide the traveling time of the tested vehicle (Tt).
range assignment (dTra) in And by NaSch-S2S model (Jost et al., 2003), they
Vehicular ad hoc networks é h +1 ù
-1
ê (1 - Ts / Tt ) ú
get the density K = ê + 1ú of
In addition to speed, density is also a common ê l¢ ú
êë úû
parameter of traffic information. Here we will
neighbor vehicles around, η and λ are given as the
introduce a method, DTRA, in vehicular network
levels of the estimated traffic condition. At least,
which can estimate local density and dynamically
although the larger the radius of the inter-vehicle
adjust transmission radius without other vehicles
transmission range, the easier will be the commu-
or road side unit (RSU) (Artimy, 2007). Accord-
nication; but on the same time, it will be easier to
ing to the definition of traffic flow, the number
be intterrupted. Thus, the area of communiction
of vehicles passing through a fixed-point per
can be adjusted according to the dynamic traffic
unit time could be computed as velocity (driving
condition mentioned above, reducing the radius
distance per unit time) product density (number
of communication in high intensity condition and
of vehicles per unit distance). But as shown in
vice versa, so that highly connected network and
Figure 14.9, traffic flow can not always reflect
data transmission rate can be achieved.
the traffic condition, such as conjunction or free
238
evaluation of a neighbor-based in its communication range, and U become the

Vehicle density estimation scheme probe vehicle doing the same work as vehicle V
in the one-hop situation, collect the data along the
Most current methods about collecting and direction and return the result back to vehicle V,
processing traffic information are still based finally the two-hop result is gotten. Evaluating the
on centralized communication model. It will be inaccuracy between estimation and actual density
delivered to every vehicle after the processing of by MAE (mean absolute error), the authors find
the central processing unit. However, this article that the two-hop estimation is the better one. And
is collected to predict the density of the vehicles the inaccuracy increases if the density is too low.
on the roads based on the inter-vehicle commu- The reason may due to the lack of vehicles in the
nications (Panichpapibo et al., 2008). It takes two communication range. In the end, the author sug-
models to test, One-hop-Neifhbor Scheme and gests three ways to improve the estimation of the
Two-hop-Neighbor-Scheme. The processes are density. Firstly, enlarge the communication range
as follows: (1) in the case of one-hop senario, to avoid the shortage of vehicles; secondly, rise
vehicle launched to send “HELLO” packet to the the hops up; last and the most effect, use more
vehicles which are in the communication range. samples for evaluation.
And other vehicles send a message back to the
collecting vehicle on receiving the packet. If the estimating Travel Times and
inter-vehicle space shows the exponential distri- Vehicle Trajectories on Freeways
bution, then the number of vehicles will show the using dual loop detectors
poisson distribution (Wisitpongphan et al., 2007).
At this time, the maximum probability of local There are many physical methods to depict traf-
density is the best expression to the global density fic jams before 2002 (Nagatani, 2002). However,
estimation ( r̂ * ) equals to the numbers of neighbor someone did not satisfied in using velocity and
vehicles (k) divide the communication radius of density as the only factor in the traffic informa-
vehicles (z). (2) In two-hop situation, vehicle V tion. This article provides a method to estimate
broadcasts inquiring packets and receive responses travel time by using Dual Loop Detetor. The
from the vehicles within the communication range. author uses the Dual Loop Detetor to record the
Then send the packet to the furthest vehicle U speed of each vehicle and its arriving time at the
Figure 9. Flow-density
239
Figure 10. Travel time estimation “a” is a chord

the certain point where the detector is set. Then
draw chords to represent the directions and loca-
tions for vehicles. Because there may be free flow
or conjunction at the same time, the situation of
conjunction into consideration need to be taken
(Figure 9), using the slop of conjunction uc to
cut off the chords (Figure 10), then sum up the
cutting chords to get the estimation of vehicles’
path and travel time.
As the traffic information develops rapidly,
the traffic information in VANETs is bound to transmission technology, automation technology,
become more important. Besides all the traffic artificial intelligence, and electronic technology
informations mentioned above, which includes etc. And establish a large-scale and all-round play
the vehicle speed, vehicle density, and travel transportation management system which is on
time, there are still other traffic information; for time, accurate, and efficient.
example, the instant safety information. But no
matter what kind of information is given, the most 14.4.1 Vehicle routing
important aim is to let the users to arrive at their navigation system
destinations safely and swiftly.
Navigation system is usually defined as the instru-
ment which can be used as guidance for routing.
Vehicle naVigaTion Technology From the early compass even later paper maps, the
recent development of handheld GPS map naviga-
In recent years, the expansion of cities scale and tion, on-board navigation systems (mainly can be
urban population increase as well as the develop- divided into two types of portable and embedded),
ment of automobile industry. Urban traffic prob- as well as PDA and Smartphone combined with
lem such as traffic congestion, traffic accidents, navigation systems. Vehicle navigation system
transport environmental degradation, and energy directs vehicle along a path from start position to
shortages has become the world facing common the destination in the road network. The system
problems. According to statistics, road congestion attempts to generate an optimal path of guidance,
has caused a great amount of direct or indirect almost all the systems try to find the shortest travel
economic losses. Thus, in order to improve ef- time for navigation. In other words, it attempts
ficiency and security of transportation, every to search for the fastest route in the road network
country has developed vehicle navigation and for guidance. The key influencing factors of the
positioning system one after another. To solve the vehicle navigation system architecture are vehicle
traffic congestion in a traditional way is to expand localization technology and the integrity of the
the roads or to construct new roads. But with the electronic map.
living environment nowadays, the urban popula- Global Positioning System is a widely used
tion decreasing, the use of traditional methods satellite localization system. There are 24 satellites
is difficult to solve traffic congestion problems. distributed over the Earth. The satellite uses basic
Undoubtedly, the modern technology comes with triangulation positioning principle, combined with
the tide of fashion to solve the traffic problem the development of satellite and communication
effectively. It integrates computer technology, technology to GPS receiver to receive GPS satellite
information technology, data communications signals by the wave. It can accurately calculate
240
the time, position, and velocity of the area where as well as centralized and decentralized system
are measured on the surface of the Earth at any- into comparison.
time and anywhere. Its advantages are global,
high precision, real-time positioning speed, can static vs. dynamic system
be used for many users simultaneously, and free
of charge. And the disadvantage is that it can not One important distinction of routing guidance
receive signals due to the signals obscured in system is whether the system uses a static or
some areas, such as behind buildings or tunnels, a dynamic approach. This characteristic of the
navigation signal interruption, therefore, must be system dictates overall operation, no matter what
again orientation. vehicles receive and react to update path informa-
The components of vehicle navigation systems tion during roadway propagation.
requirement: (1) positioning accuracy, combin- A static routing algorithm determines and main-
ing different navigation technology to improve tains a path for a user from the source to destination
positioning accuracy. (2) Complete the correct regardless of changing the traffic characteristics in
geographic information systems, to update map time. In this system, vehicles already on the road
data and road network instantly, and continued with assigned paths do not react to any real-time
to increase the road property and other relevant network changes. Clearly this system represents an
information. (3) Friendly user interface. (4) Easily offline routing approach. Several papers investigate
to operate, concern about the development of voice the idea of static path route selection such as in
recognition capabilities, to provide voice-guided (Dijkstra, 1959; Dreyfus, 1969).
Service. (5) Price considerations. On the other hand, dynamic path planning
Information displayed can be divided into two incorporates real-time traffic information and
categories, (1) Driving geographic information: to reacts to the changing conditions of the roadway.
provide users with a wealth of traffic information, This allows updating of the path of a vehicle as
such as gas stations, parking area, restaurants, the user proceeds to the destingtion. This type
amusement parks, financial institutions etc. Most of routing provides a more robust solution than
of the current systems are well build, and the con- static routing for congestion alleviation. The dif-
tent of the database is increase continuing. And ferences between static and dynamic navigation
the parts of the systems also provide users with system are system design and requirements. Static
custom landmark features. (2) Driving guidance routing guidance systems require less roadway
information: to get the real-time traffic condition infrastructure and computing power, the system
information, such as traffic congestions, traffic ac- is vulnerable to events and congestion that occurs
cidents, construction or temporary traffic control on the roadway. Dynamic routing guidance is a
measures etc from communications technology. robust operation during events and congestion, but
Users can avoid road congestion or accident lo- the system required more computing power and
cation timely, and select one alternative path for roadway infrastructure. Several papers investigate
their route planning. the idea of dynamic path route selection such as
Based on the above classification, comment, in (Hall, 1986; Karimi et al., 2004).
and comparison of vehicle navigation system
today (Schmitt et al., 2006), it will be divided deterministic vs. stochastic system
into four main categories investigate separately
for static and dynamic systems, deterministic and The distinction between these two systems is
stochastic system, reactive and predictive system, mainly due to treating the cost of traveling from the
241
origin to destination as a deterministic or stochastic at the cost of computational complexity. Stochastic

cost. This cost is usually the travel time. routing systems clearly require a greater amount
Deterministic routing systems assume deter- of data for routing guidance. Thus, the data re-
ministic parameters for roadway links without quirement, especially historical data, increases
regard to any random nature of traffic. These computational time and data storage costs.
parameters may be updated frequently in time.
The most commonly used metric for deterministic reactive vs. predictive system
routing is the distance traversed by a vehicle along
the route. Calculation of the path using a deter- This classification is very important. It controls
ministic approach would yield the total distance. the complexity, robustness, and vulnerability of
The total travel time is then calculated using mean events. The main differences between the two
vehicle speed without considering the influence systems are their complexity and reliability.
of any variation due to the stochastic nature of Reactive navigation routing is also a kind of
the traffic. Several papers investigate the idea of feedback routing, which based on the recent situ-
deterministic path of routing selection such as in ation of the travel network. The routing system
(Friesz et al., 1989; Zhan, 1997). determines the paths by real-time traffic informa-
Stochastic routing uses a more complex ap- tion, but will not predict the coming situation.
proach to vehicle routing in which the random This system will provide the best route, which
nature of the traffic condition is also considered. having least travel time, to the users. The system
In this approach, traffic along the road may be has a less complexity, but a greater vulnerability
considered either as a stationary or non-stationary because of road congestion and unexpected events.
stochastic variable. Stationary stochastic routing Several papers investigate the idea of reactive
guides vehicles along a path determined using only path route selection such as in (Pavlis et al., 1999;
mathematical expectation of conditions deemed Deflorio, 2003).
imporant by the algorithm. These conditions are Predictability routing of navigation can es-
assumed to be time invariant; therefore, they are timate the routes. It is based on the prediction
not changing with time. In most cases of vehicle models from the predicted situations of the future.
routing the mean and variance of travel time are The Predictive System using a repeated model of
used. Several papers investigate the idea of sta- traffic information and useful historic information
tionary stochastic path route selection such as in to predict the coming situations, which is based
(Bertsekas et al, 1991; Alexopoulos, 1997). Non- on linking or the region of network. The system
stationary stochastic routing takes into account the will provide the best route with least travel time
changing characteristics of time with respect to the according to the estimated traffic conditions.
traffic condition. These changing conditions are Several papers investigate the idea of predictive
familiar with mean travel time and the variance path route selection such as in (Messmer et al.,
of the function which is time in one day. Several 1990; Kotsialos et al., 2002).
papers investigate the idea of stationary stochastic Although the reactivity and the predictability
path route selection such as in (Seongmoon et al., routing are working in different ways, they both
2005; Fu, 2001). provide the best route to the users. We can say
Deterministic systems offer the simplest solu- that the reactive system can be executed in the
tions while stochastic non-stationary solutions environment of lower cost and less complexity,
offers the most robust answer to the problem and the prediction system will be able to avoid
capturing the charateristics of traffic congestion, the generation of traffic congestion.
242
centralized vs. decentralized system In order to expand the commercialization of the

product and to accelerate it into human life from
Routing navigation can be dominated by a cen- technology products converted to daily necessities,
tralized system and monitoring dynamics of the related industries should study in market prefer-
entire network, or optimize the performance of a ence actively when technology developed.
single user himself.
Centralized system is focus on providing
routing selections, and the distributed system daTa disseminaTion in VaneTs
provides personalized real-time road informa-
tion selections. Centralized routing system uses data dissemination
a central unit to provide the routing selection.
The system uses a predictive or reactive routing Data dissemination concerns about the transmis-
to guide vehicles for the best route to the destina- sion of information to intended receivers with
tion in the network. Centralized routing is more certain design objectives. The design objectives
complex than distributed system. This method that considered include low delay, high reliability,
provides more robust and reliable system to all low memory occupancy, and low message passing
hierarchies of the Internet. Centralized system will overhead (Nadeem et al., 2006).
provide better network capacity to the network, Four services that have immediate application
the best of the road. It will calculate and deal with are unicast, multicast, anycast and scan. “Unicast
the road side infrastructures. Another concerned with precise location” means a message should be
issue of the centralized system is the compliance delivered to node i in location l before time t . “Uni-
rate of Internet users. If the instructions do not cast with approximate location” means sending a
be followed by the drivers, the effectiveness of message to node i before time t1 while that node
navigation systems will reduce while compliance was last known to be at location l with mobility
rate decreasing. Several papers investigate the m at time t2. “Multicast” means disseminating a
idea of centralized path route selection such as in message to all receivers in region r before time t.
(Yamashita et al., 2004 ; Tomkewitsch, 1991). “Anycast” means disseminating a message to one
Distributed routing navigation system required among a set of possible destinations in region r
for individual users to determine all the routes before time t. “Scan” is to have a message traverse
of navigation, and each one can operate by the region r once before time t.
machine or man-made to decide the route of Data delivery mechanisms define the rules
navigation. Road infrastructures can not control for moving information through the network, for
the selected routes of users, but it can provide example “Node centric”, “Location centric”, “Op-
current information and the parameters which portunistic forwarding”, and “Trajectory based
has predicted of the network. The system may forwarding” .The “Node centric” approach speci-
not be robust on the Internet hierarchy, because fies the routing path as a sequence of connected
the optimization is for the situation of each user nodes. However the high vehicle mobility in V2V
rather than to whole network. The critical differ- networks will quickly render inter-node connec-
ence of centralized and distributed routing navi- tions invalid. The “Location centric” approach
gation system is that the computation for users decouples the routing path from the intermediate
or infrastructures. Several papers investigate the nodes and the message is forwarded to the next
idea of decentralized path route selection such hop(s) closer to the destination geographically.
as in (Wunderlich et al., 2000 ; Minciardi et al., If a hole is encountered, efforts are made to find
2001). a path around it. “Opportunistic forwarding,” as
243
suggested in, targets networks where an end-to- forwarding trajectory, etc., is specified by the data
end path cannot be assumed to exist. Messages source and is placed in the message header.
are stored and forwarded as opportunities present A forwarding trajectory is specified as a path
themselves. When a message is forwarded to an- extending from the source to the destination re-
other node, a copy may remain with the original gion. One of the MDDV objectives is to deliver
and be forwarded again later to improve reliability. messages to their destination regions as soon as
“Trajectory based forwarding” directs messages possible. A naive approach would be taking the
along predefined trajectories. It was presented to path with the shortest distance from the source
work well in a dense network. Despite their sparse- to the destination region. However, information
ness, V2V networks should be a natural application propagation along a road depends largely on the
of trajectory based forwarding because messages vehicle traffic on it. A short road distance may not
are moving along the “road” graph. translate to short information propagation delay.
High vehicle density often leads to fast informa-
mobility-centric data dissemination tion propagation. But vehicle traffic conditions
algorithm for Vehicular network change over time and vary from one road segment
to another. Here, the authors only explore the
MDDV is a “mobility centric” approach that static road network topology information since
combines opportunistic forwarding, geographi- road networks are typically engineered to match
cal forwarding, and trajectory based forwarding. transportation demands. They define d(A,B) as the
A forwarding trajectory is specified extending “Dissemination length” of a road segment from
from the source to the destination (trajectory road node A to B, which takes into account the
base forwarding), along which a message will be static road information.
moved geographically closer to the destination
(geographical forwarding). With an opportunistic d(A,B) = r(A,B) (m-(m-1)(ip+cjp)), 0< c < 1
forwarding approach, rules must be defined to
determine which is eligible to forward a message,
when a copy of the message should be passed to where, r(A,B): road length from A to B
another vehicle, and when a vehicle should hold/
drop a message. They motivate the design by ref- i: number of lanes from A to B
erence to a test scenario, geographical-temporal
multicast. Geographical-temporal multicast is j: number of lanes from B to A
formally defined as: deliver a message to all
vehicles in/entering region t before time t while But the traffic in the opposite direction of the
the data source s is outside of r. desired information flow is less helpful than the
Assume that a vehicle knows the road topol- traffic in the same direction of the information flow.
ogy through a digital map and its own location in Constant c is used to discount the opposite traffic
the road network via a GPS device. And suppose flow. In the reference (Wu, Fujimoto, Guensler,
vehicles know the existence of their neighbors & Hunter, 2004), when i =1, j =0, d(A,B)=r(A,B),
through some link level mechanism. But they do then set m=5, p=0.1, c=0.005.
not assume a vehicle knows the location of its The dissemination process consists of two
neighbors. The message dissemination informa- phases: the forwarding phase and propagation
tion, e.g., source id, source location, generation phase. In the forwarding phase, the message is
time, destination region, expiration time and forwarded along the forwarding trajectory to the
destination region. When the message reaches the
244
destination region, the propagation phase runs and a non-message head candidate can only learn from
the message is propagated to every vehicle in an received messages.
area centered on the destination region before the There are rules for a message holder to transit
message time terminated. During the forwarding between a message head candidate and non-
phase, the authors call the “message head” that message head candidate. Suppose the current
the message holder closest to the destination re- time is tc, a vehicle’s current location is lc, and
gion along the forwarding trajectory. The vehicle a vehicle’s installed message head pair is l, t
taking the role of the message head may change where l is the message head location and t is the
over time as the message propagates. With perfect generation time.
knowledge, every vehicle knows the message head
vehicle in real time. When the message head tries 1. Non-message head candidate ⇒ message
to pass the message to other vehicles that may head candidate: During the forwarding
be closer to the destination region. During the phase, one important observation is that a
propagation phase the message is propagated to vehicle passing its installed message location
vehicles without the message in the specified area. in a shorter period after the generation time is
With the assumption that vehicles do not know the more likely to be the message head because
location of others, this is difficult to do. after a long period the message may have
In both cases, the message is lost. To overcome already been forwarded far away toward the
this problem, they allow a group of vehicles near destination region along the trajectory. Thus,
the real message head to actively forward the mes- they stipulate that a non-message head can-
sage instead of the message head vehicle only. didate becomes a message head candidate if
The message head pair, the message head it passes its installed message head location
location and its generation time, is contained to toward the destination region before t + T1,
the message. The “message head pair” provides where T1 is a system parameter. During the
the more correct knowledge of a message holder propagation phase, message holders moving
concerning the message head location. The actual into the destination region assume the role
message head can move either toward or away of the message head candidate.
from the destination region along the forwarding 2. Message head candidate ⇒ non-message
trajectory within a short period of time. But it head candidate: During the forwarding
should move toward the destination region in the phase, there are two transition rules:
long term. For simplicity, the authors require that 1. If the message head candidate leaves
the message head location installed by a message the trajectory or moves away from the
holder never moves backward, which means that destination region along the trajec-
a message holder can only install a new message tory, it becomes a non-message head
head location closer to the destination region candidate;
than the one currently installed. To reduce the 2. If a message head candidate moves
publication and dissemination of false informa- toward the destination region along the
tion, only some vehicles are allowed to generate trajectory, it stays as a message head
the message head pair. A message holder may candidate until it receives the same
assume either one of two roles: the message head message with another message head
candidate and non-message head candidate. Only pair l n , tn where ln is closer to the
a message head candidate can actively publish its destination region than lc.
current location as the message head location, and
245
During the propagation phase, a message head state, tc < t + T3, and lc between L3 with l.
candidate becomes a non-message head candidate T2, T3, L2, L3 are system parameters and T2
once it moves out of the destination region. < T3, L2 < L3.
A message holder updates its installed message 2. Propagation phase: A message holder can
head pair with the information from received mes- be either in the active state or not eligible to
sages. Two messages differing only in the message transmit. A message holder is in the active
head pair are two versions of the same message. state if tc < t + T2 and lc is within the distance
One message version with message head pair L2 from l.
li , ti is said to be newer than another message
version with message head pair l j , t j An opportunistic forwarding mechanism must
determine when to store/drop a message. The
1. li is closer to the destination region than lj design decision can affect delivery reliability,
2. li = lj but ti > tj memory usage and message overhead. The deci-
sion to store/drop messages can be based on a
Data exchange can be triggered by several vehicle’s knowledge of its future movement tra-
types of events. In MDDV, data exchange is trig- jectory. For example, assume vehicles are aware
gered by: new messages, newer message versions of its own near future movement trajectory, a
or older message versions are received, or new message holder may decide to drop a message
neighbors appear. Transmissions triggered by if it knows that continually holding the message
new messages or newer message versions serve can no longer contribute to suppress unneces-
to quickly propagate messages or dissemination sary message transmissions based on its future
status. Transmissions triggered by older message movement trajectory. In MDDV every vehicle
versions can help eliminate false/obsolete informa- stores whatever it overhears since this is almost
tion. This scheme has both the advantages of fast free except occupying memory buffers. A vehicle
delivery and high delivery reliability. It is called drops a message when the vehicle leaves the pas-
the “full protocol.” sive state during the forwarding phase, leaves the
The data exchange algorithm is defined as: active state during the propagation phase or the
message expiration time elapses.
1. Forwarding phase: A message holder can
be either one of following dissemination concept of replication
states: the active state, the passive state,
and not eligible to transmit at all. A mes- In the Intelligent Transportation System (ITS), the
sage holder in the active state runs the full car to car wireless network and car’s communica-
protocol to actively propagate the message tion to the infrastructure (C2X communication)
while a message holder in the passive state have been given high priority recently. This trend
only transmits the message if it hears some is corresponding to use basis of the C2X active
older message version. The active propaga- safety applications to reduce the traffic accident
tion can help populate the message, move or other seriously expected.
the message closer to the destination region The networks constructed by C2X is called
or update dissemination status. The passive “Vehicular Ad Hoc Networks (VANETS)”, it
updating serves to eliminate false/obsolete runs distributed approach and does not depend on
information only. In the active state, tc < t + the existing infrastructure to operate. Because of
T2, and lc between L2 with l. In the passive the C2X node is generally moving, so the whole
246
network is divided by a number of Ad Hoc net- This method keeps storing their current and
work composed. recent location information in a period of time.
Due to the region information providing an Therefore vehicle can accurate to provide their
appropriate way to describe that region, its ap- tracking over the past. It can be every kind of
plications have various C2X’s characteristics. straight-line (ex: driveway) or curved-line (ex:
For example, in the case of traffic jam alert, the driveway exit or highway). If the slippery road
main of provision will directly affect the situa- warning occurs, it is necessary to send the warn-
tion of information to the vehicles, as vehicles ing message to the vehicle. Thus, the position
have nearly the same lane. Other major factors information of that area must be stored actually.
include the communication between vehicles The saving points are probably in hundred meters
and infrastructure. To imagine a case, if the traf- away. All settings can be illustrated that greatly
fic light can broadcast the current signal to the reduce digital map by the regional road network
nearest car. In this case where the vehicles close in the travel path of vehicles to access. To decide
to the traffic signal, it can specify the affected the location of information storage and the use of
roads and driving directions. However, this as- the map are very easy problems, because most of
signment is hard to achieve in the reality traffic the locations are dependent on the technology of
situation, because the moving nodes may already current trends (Roessler et al., 2006).
get messages in hundred meters away and roads Therefore, it contains the following features:
are generally not linear. Therefore, to estimate the
information received by vehicles is a challenging • Form of the Roads: The form of different
problem. This article introduce about how to deal roads will be different number of support-
with the roads affected that allow changing and ing points. If there is lots of curve in the
ensure the problem that vehicles can receive the road, it is necessary to rebuild the storage
relevant information. location information. Otherwise, it only
needs two points of information while the
replication concept of road is straight.
regional road network • Velocity of the Vehicle: When vehicle
moves with high speed in a short time in-
The approaches in real traffic environment de- terval, it generally store high change loca-
scribing the region are often unrealistic, because tion information.
it ignore the specific situation of the driving
environment. In the affected region may be more To go along with the interpolation points,
complicated to operate and not at all the nodes in the width of the addressing region (corridor), is
this area will participate in. In fact, the relationship needed. This width can be equal to the whole set
about the car direction is an important problem. of supporting points in order to reduce the amount
The issue of concept is to use some of the chal- of data. However, it might also be necessary to
lenges faced the basic position of a temporary specify different width parameters in case of
mechanism to change the essential part of the a complicated road form. Figure 11 shows the
street. This means that each vehicle contains C2X example that is from X1/Y1,…,Xn / Yn of support-
are required a wireless communication compo- ing points.
nents and similar Global Position System (GPS) It can be found that the distance between two
location systems. However, not every vehicle points are different from the whole addressing
requires Global Position System. region. To collect this kind of additional infor-
mation from supporting points, i.e. the expected
247
Figure 11. Example for concept of replication
information of driving directions which is form the station like a traffic sign is the sender, the
addressing area linked to a dedicated road section. position information may alternatively be
This new method includes several advantages that hard-coded in the device.
are summarized below:
As mentioned above, the addressing approach
• Low Basic Hardware Necessary: Whole is not restricted to the vehicle to vehicle scenario
information needed like current position, that has been used to describe the technique. It can
driving direction and time can be got from be further applied in different C2X-applications
a basic positioning system. However, the like the communication between a car and a traf-
memory demand can be constraint by limi- fic signal system.
tation of the maximum number of cached
position data.
• Simple Relevance Evaluation: Using reFerences
the received information vehicle to make
a first evaluation of the relationship of the Alexopoulos, C. (1997). State Space Partitioning
information even if they are not equipped Methods for Stochastic Shortest Path Problems.
with a navigation system. Networks, 1(30), 9–21. doi:10.1002/(SICI)1097-
• Scalability: There are several method 0037(199708)30:1<9::AID-NET2>3.0.CO;2-H
to improve. For example, if a vehicle is Artimy, M. (2007). Local Density Estimation and
equipped a navigation system, it may also Dynamic Transmission-Range Assignment in
solve the problem that address streets were Vehicular Ad hoc Networks. IEEE Transactions
not on its recent route. But it might also be on ITS, 18(3), 400–412.
affected.
• Flexible: This approach can apply to any
mobile and fixed nodes. If an immobile
248
Bertsekas, D. P., & Tsitsiklis, J. N. (1991). An Karimi, A., Hegyi, A., Schutter, B. D., Hel-
Analysis of Stochastic Shortest Path Problems. lendoorn, J., & Middelham, F. (2004, Oct.).
Mathematics of Operations Research, 16(3), Integrated model predictive control of dynamic
580–595. doi:10.1287/moor.16.3.580 route guidance information systems and ramp
metering. Paper presented at the Proceedings
Deflorio, F. P. (2003). Evaluation of a reactive
of the 7th International IEEE Conference on
dynamic route guidance strategy. Journal of the
Intelligent Transportation Systems (ITSC 2004),
Transportation Research Board Transportation,
Washington, DC.
11(5), 375–388.
Kastinaki, V., Zervakis, V., & Kalaitzakis, K.
Dijkstra, E. W. (1959). A note on two problems in
(2003). A survey of video processing techniques
connection with graphs. Nuerische Mathematik,
for traffic applications. Image and Vision Com-
1, 269–271. doi:10.1007/BF01386390
puting, 21(4), 359–381. doi:10.1016/S0262-
Dreyfus, S. E. (1969). An Appraisal of Some 8856(03)00004-0
Shortest Path Algorithms. Operations Research,
Kotsialos, A., Papageorgiou, M., Diakaki, C.,
17, 395–412. doi:10.1287/opre.17.3.395
Pavlis, Y., & Middelham, F. (2002). Traffic flow
Friesz, T. L., Luque, J., Tobin, R. L., & Byung- modeling of large-scale motorway networks us-
Wook, W. (1989). Dynamic network traffic as- ing the macroscopic modeling tool METANET.
signment considered as a continuous time optimal Journal of IEEE Transaction on Intelligent Trans-
control problem. Operations Research, 37(6), portation Systems, 3(4), 282–292. doi:10.1109/
893–901. doi:10.1287/opre.37.6.893 TITS.2002.806804
Fu, L. (2001). Adaptive Routing Algorithm for Messmer, & Papageorgiou, M. (1990). META-
In-Vehicle Route Guidance Systems with Real- NET: A macroscopic simulation program for mo-
Time Information. Transportation Research Part torway networks. Journal of Traffic Engineering
B: Methodological, 35(8), 749–765. doi:10.1016/ & Control, 31(8-9), 466-470.
S0191-2615(00)00019-9
Minciardi, R., & Gaetani, F. (2001). A decentral-
Hall, R. (1986). The Fastest Path through a ized optimal control scheme for route guidance
Network with Random Time- Dependent Travel in urban road networks. Paper presented at the
Time. Transportation Science, 20(3), 182–188. Proceeding of IEEE Intelligent Transportation
doi:10.1287/trsc.20.3.182 Systems, Oakland, CA.
Hudson Valley Transportation Management Cen- Nadeem, T., Shankarand, P., & Iftode, L. (2006,
ter (HVTMC) (n.d.). Retrieved from http://www. July). A comparative study of data dissemination
hudsonvalleytraveler.com/Glossary.html models for vanets. Paper presented at the Pro-
ceeding of 3rd Annual International Conference
Jost, D., & Nagel, K. (2003). Probabilistic traf-
on Mobile and Ubiquitous Systems (MOBIQUI-
fic flow breakdown in stochastic car following
TOUS), San Jose, CA.
models. Transportation Research Record, 1852,
152–158. doi:10.3141/1852-19 Nagatani, T. (2002). The physics of traffic jams.
Reports on Progress in Physics, 65(9), 1331–1386.
doi:10.1088/0034-4885/65/9/203
249
Noble, B., Yoon, J., & Liu, M. (2007, June 11-14). Ryu, B., Yeung, G., Krishnan, H., Yin, J., ElBatt,
Surface Street Traffic Estimation. Paper presented T., Habermas, S., et al. (2004). Performance Evalu-
at the Proceedings of International conference on ation of Safety Applications over DSRC Vehicular
Mobile systems, San Juan, Puerto Rico. Ad Hoc Networks. Paper presented at the Proceed-
ings of the 1st ACM international workshop on
Panichpapibo, S., & Pattara-atikom, W. (2008,
Vehicular ad hoc networks, Philadelphia, PA.
Oct.). Evaluation of a neighbor-based vehicle
density estimation scheme. Paper presented at Schmitt, E. J., & Jula, H. (2006, Sep.). Vehicle
the ITST 2008. 8th International Conference, Route Guidance Systems: classification and
Phuket. Comparison. Paper presented at the Proceedings
of the IEEE Intelligent Transportation Systems
Pavlis, Y., & Papageorgiou, M. (1999). Simple De-
Conference (ITSC2006), Toronto.
centralized Feedback Strategies for Route Guid-
ance in Traffic Networks. Transportation Science, Seongmoon, K., & Lewis, M. E., & III, C. C. W.
33(3), 264–278. doi:10.1287/trsc.33.3.264 (2005). Optimal vehicle routing with real-time
traffic information. IEEE Transactions on Intel-
Real-Time Traffic Network Systems of Cities and
ligent Transportation Systems, 6(2), 178–188.
Counties in Taiwan. (n.d.). Retrieved from http://
doi:10.1109/TITS.2005.848362
www.e-traffic.com.tw
Sichitiu, M. L., & Kihl, M. (2008). Inter-Vehicle
Rice, J., & Cho, Y. (2006). Estimating Veloc-
Communication Systems: A Survey. IEEE Com-
ity Fields on a Freeway From Low-Resolution
munications Surveys & Tutorials, 10(2), 88–105.
Videos. IEEE Transactions on Intelligent Trans-
doi:10.1109/COMST.2008.4564481
portation Systems, 7(4), 463–469. doi:10.1109/
TITS.2006.883934 Taleb, T., Sakhaee, E., Jamalipour, A., Hashimoto,
K., Kato, N., & Nemoto, Y. (2007). A Stable Rout-
Robert, H., & Ilya, P. (1979). A Two-Fluid Ap-
ing Protocol to Support ITS Services in VANET
proach to Town Traffic. Science, 204(4389),
Networks. IEEE Transactions on Vehicular
148–151. doi:10.1126/science.204.4389.148
Technology, 56(6), 3337–3347. doi:10.1109/
Robinson, C. L., Caveney, D., Laberteaux, K., TVT.2007.906873
& Caminiti, L. (2006, September 29). Efficient
Tomkewitsch, R. v. (1991). Dynamic route
Coordination and Transmission of Data for Co-
guidance and interactive transport management
operative Vehicular Safety Applications. Paper
with ALI-SCOUT. Journal of IEEE Transac-
presented at the Proceedings of the 3rd interna-
tions on Vehicular Technology, 40(1), 45–50.
tional workshop on Vehicular ad hoc networks,
doi:10.1109/25.69971
Los Angeles, California, USA.
Wisitpongphan, N., Bai, F., Mudalige, P., Sadekar,
Roessler, B., & Bruns, C. (2006, Mar.). Replication
V., & Tonguz, O. (2007). Routing in sparse ve-
of Local Road Networks for Regional Information
hicular ad hoc wireless networks. IEEE Journal
Dissemination in VANETS. Paper presented at the
on Selected Areas in Communications, 25(8),
Proceedings of the 3rd Workshop on positioning,
1538–1556. doi:10.1109/JSAC.2007.071005
Navigation And Communication (WPNC’06),
University of Hannover, Germany.
250
Wu, H., Fujimoto, R., Guensler, R., & Hunter, Yamashita, T., Izumi, K., & Kurumatani, K.
M. (2004, Oct.). Mddv: a mobility-centric data (2004). Car navigation route information shar-
dissemination algorithm for vehicular networks. ing for improvement of traffic efficiency. Paper
Paper presented at the Proceedings of ACM Ve- presented at the Proceeding of 7th international
hicular ad hoc networks (VANET), Philadelphia, IEEE conference on Intelligent Transportation
PA. Systems, Washington, D.C., USA.
Wunderlich, K. E., Kaufman, D. E., & Smith, R. Zhan, F. B. (1997). Three fastest shortest path
L. (2000). Link travel time prediction for decen- algorithms on real road network: data structure
tralized route guidance architectures. Journal of and procedures. Journal of geographic informa-
IEEE Transaction on Intelligent Transportation tion and decision analysis, 1(1), 69-82.
Systems, 1(1), 4–14. doi:10.1109/6979.869017
251
252
Chapter 15
DTN Technologies for
Vehicular Networks
Kun-Chan Lan
absTracT
A Delay Tolerant Network (DTN) is one type of challenged network where network contacts are inter-
mittent or link performance is highly variable or extreme. In such a network, a complete path does not
exist from source to destination for most of the time. In addition, the path can be highly unstable and
may change or break unexpectedly. To make communication possible in a delay tolerant network, the
intermediate nodes need to take custody of data during the blackout and forward it toward the desti-
nation when the connectivity resumes. A vehicular network nicely falls into the context of DTN since
the mobility of vehicles constantly causes the disruption of link connectivity’s between vehicles. In this
chapter, the authors discuss some research challenges and issues which might occur in a Delay Tolerant
Network and how they are related to vehicular networks.
inTroducTion installations. Due to node mobility or limited radio

range, end-to-end paths do not always exist between
Delay/disruption tolerant networks (DTN) can be some or all the nodes. As a result, data are transferred
used to increase the robustness of the network where in a store-carry and forward paradigm. Vehicular
the network protocols must be explicitly designed ad hoc networks (VANET) have been envisioned
to perform despite frequent disruptions. A typical to be useful in road safety and many commercial
DTN consists of a set of wireless nodes, some or applications. For example, a vehicular network can
all of which are mobile. The nodes can range from be used to alert drivers to potential traffic jams,
small sensor nodes, to mid-sized devices carried by providing increased convenience and efficiency. It
people, robots or vehicles, to larger semi-permanent can also be used to propagate emergency warning
to drivers behind a vehicle (or incident) to avoid
DOI: 10.4018/978-1-60566-840-6.ch015 multi-car collisions. To realize this vision, the FCC
DTN Technologies for Vehicular Networks
has allocated 75 MHz of spectrum for dedicated stack. At the physical layer, directional antenna/
short range communications (vehicle-vehicle or MIMO techniques may be applied to increase
vehicle-roadside), and the IEEE is working on the network capacity. At the data link layer,
standard specifications for inter-vehicle com- new MAC protocols should be designed to meet
munication. latency and reliability requirements, especially
As technology advances make it more feasible for safety related applications. Because vehicles
and cost-effective to produce vehicles that are move along roads, directional-antenna-based
equipped with communication capabilities that MAC mechanisms might be especially useful for
allow for inter-vehicle communication, large VANET. At the network layer, protocols should
scale vehicular ad hoc networks are expected to be designed to exploit the mobility to maintain
be available in the near future. There are various the route. At the transport layer, new protocols
kinds of VANETs based on the entities involved, should be designed to tolerate routing layer dis-
such as vehicle to vehicle communication (V2V), ruptions. From the network perspective, security
vehicle to roadside communication (V2R) or ve- and scalability are two significant challenges.
hicle infrastructure integration (VII), and roadside Although power efficiency in VANET is less of
to roadside communication (R2R). a concern, scalability may still be critical. Due
There are different ways to look at DTN and to the nature of the vehicular applications, there
VANET. At one extreme, DTN is more general, might be more flooding/broadcasting in VANET
and VANET is a special kind of DTN. At another than in traditional ad hoc networks.
extreme, DTN techniques are only used in some This could easily create network congestion
VANET applications when vehicles are far away if the communication protocols are not well-
from each other, but not in safety related ap- designed.
plications which have strict delay requirements. There are many challenges and opportunities
Nevertheless, DTN and VANET have many com- in DTNs. Many VANET applications can be built
mon characteristics which present challenges and on DTNs, and hence most of the challenges and
opportunities for the research community. opportunities presented in VANET are also valid
Many unique characteristics of VANET bring in DTNs. As social networking becomes popular,
out new research challenges. First, due to fast more applications can be built over DTNs. Also,
vehicle movement, network topology and chan- a number of non-interactive applications related
nel conditions change rapidly. As a result, many to sensor networks can be adapted to DTNs.
well-studied structures such as tree, clustering, Unlike Internet applications, DTNs provide an
grid, are extremely hard to set up and maintain. opportunity for aggregation, replication, and in-
Second, the network density is highly dynamic. network processing as data traverses the network.
The traffic load is low in rural areas and during Traditionally, DTN is treated as an edge solution
night, which may result in frequent disconnections for the future Internet paradigm. It is a challenge
and network partitions. to re-think the Internet as a DTN. Further, DTNs
On the other hand, during rush hours or traf- may rely on some level of infrastructure support
fic congestion, the network density is very high, to improve their performance, and hence hybrid
which may generate data collisions and result in DTN/infrastructure networks are another area to
network congestion. Third, the vehicle mobility explore.
is partially predictable since it is limited by the There are many technical issues to explore in
traffic pattern and the road layout. DTNs. First, how to achieve service discovery,
These unique characteristics bring out re- context awareness and group communication,
search issues at different layers of the network which are standard modules in distributed sys-
253
tems. Due to inherent delays in DTNs, these long or variable delay and high error rates. For
popular services need to be redesigned. Second, the intermittent connectivity, an end-to-end path
we should define the communication paradigm between the source and the destination may not
in DTN. In traditional networks, the end-to-end exist. Therefore, TCP/IP protocol will break in
concept is used. In DTN, many issues need further this kind of environment because an end-to-end
investigation. For example, what are the network path between source and destination may only
interfaces? Does DTN have a flow concept? Third, exist for brief and unpredictable periods of time.
DTN is based on the idea of store-carry and for- Furthermore, in DTN, there are long propagation
ward. Although forwarding has been well studied and variable queuing delays. Many Internet proto-
in traditional networks, we don’t have enough cols which are designed assuming quick return of
knowledge on the role of storage and it is not clear acknowledgements and data fail to work in such
how to manage storage in DTNs. Finally, there are networks. The problems of DTN as mentioned
many other research issues such as security, the above are addressed by adopting store-carry-
role of network coding for DTNs, and network forward message switching. Here entire chunks
characterization (based on e.g. Graph Theory). of message are transferred from one storage place
One key feature of VANET/DTN is the mo- to a storage place in another node along a path
bility model, which is very important for future that is expected to reach the destination.
research in this area. In VANET, the mobility There are many applications for DTN. For
model can be based on previous work done by re- particular field, this kind of applications is used
searchers in other fields such as civil engineering. in the special environment that is to be tolerant of
However, more work should be done on porting long delay and high error rate. The Interplanetary
these results to our community. This can be in Internet project must encompass both terrestrial
the form of adding mobility models for VANET and interplanetary links. Sami Network Con-
to well known simulation platforms such as ns2. nectivity (SNC) Project (Lindgren et al., 2007)
For general DTN, the mobility model is not that focuses on establishing Internet communication
clear, and is worth further investigation. Further, for Sami population of reindeer herders who live
communication and data traffic models may be in remote areas. DARPA (Krotkov, et al., 1999)
application-specific. is researching and developing the capability to
In summary, VANET/DTN are well outside perform urban reconnaissance with teams of
traditional networking assumptions and challenge small, low-cost, semiautonomous mobile robots.
our thinking about future Internet designs. In For developing regions networks, the networks
this chapter, we present an overview of existing of developing countries or the outlying area are
research activities in DTN. not usually perfect, so the foreign networks of a
lot of areas are being developed, even there is no
network at all. In (Brewer et al., 2005), several
dTn applications are developing for this kind of region
from education to health care and government
A Delay Tolerant Network (DTN) is a develop- services. This project is supported by (Tier, 2009).
ing system that suffers frequent and long dura- For environmental monitoring, this application is
tion network partitions. It is an overlay on top to monitor the animals or environment. In (The
of multiple network partitions, which may be zebranet wildlife tracker, 2009), this project is to
challenged by limitations such as intermittent track the Wildlife Tracker.
and possibly unpredictable loss of connectivity,
254
Figure 1. The DTN protocol stack

or more DTN regions and may optionally be a
host, so it must have persistent storage and support
custody transfers.
challenges
In a DTN, when nodes move away or turn off their

power to conserve energy, links may be disrupted
or shut down periodically. These events result in
intermittent connectivity. When no paths exist
between source and destination, the network parti-
tion occurs. Therefore, nodes need to communicate
with each other via opportunistic contacts through
store-carry-forward operation. In this section, we
consider three specific challenges in an opportu-
nistic network: the contact opportunity, the node
an oVerView oF delay storage and the power conserving.
ToleranT neTworks
contact
architecture
Due to the node mobility or the dynamics of
In a DTN, a network is typically separated into wireless channel, a node might make contact
several network partitions which called regions. with other nodes at an unpredicted time. Since
Traditional applications are not suitable for this kind contacts between nodes are hardly predictable,
of environment because they normally assume that they must exploited opportunistically for exchang-
the end-to-end connection must exist from source ing messages between some nodes that can move
to destination. The DTN enables the devices in between remote fragments of the network. Burns
different regions to interconnect by operating mes- et. al. (2005) classified the routing methods for
sage in store-carry-forward message fashion. The DTN based on characteristics of participants’
intermediate nodes implement store-carry-forward movement patterns. The patterns are classified
message switching mechanism by overlaying a new according to two independent properties: their
protocol layer, called the bundle layer (Scott et al., inherent structure and their adaptiveness to the
2007 ; Segui et al., 2006). The DTN protocol stack demand in the network. Other approaches pro-
is as shown in Figure 15.1. In a DTN, each node is pose message ferries to provide communication
an entity with a bundle layer which can act a host, service for nodes in the deployment areas (Zhao
router or gateway. When node acts a router, the et al., 2004, May ; Zhang et al., 2007 ; Chen et
bundle layer of router requires persistent storage al., 2007 ; Tariq et al., 2006). In addition, the
to queue bundles and forwards the bundles within contact capacity needs to be considered. In other
a single region. On the other hand, the bundle layer words, how much data can be transferred between
of gateway is not only provision the store-carry- two nodes when two nodes contact? Hui et. al.
forward mechanism, but provision the conversion (Chaintreau et al., 2005) defined two parameters,
between the lower-layer protocols of the regions contact duration and inter-contact time, that are
they span, as shown in Figure 15.1. In other words, the important factors in determining the capacity
a DTN gateway can forward bundles between two of the networks.
255
storage between contacts with other nodes. Depending

on the statistical information, a node may expect
As the described above, to avoid dropping packets, to the waiting time before the contact. However,
the intermediate nodes are require to have enough those methods may be applied to the network that
storage to store all messages for unpredictable followed a regular schedule. In (Jun et al., 2006),
periods of time until next contact. In other words, the authors propose a hierarchical power manage-
the required storage space increases a function of ment in DTN. The energy in the architecture can
the number of messages in the network. There- be conserved by using lower-power, long-range,
fore, the routing and replication strategies must radio to discover the contacts and then wake the
take the storage constraint into consideration. high-power, short-range, radio to prepare the
Vahdat and Becker (Vahdat et al., 2000) used data transmission. However, these architecture
Epidemic Routing by flooding the network to may be inefficient because mobile node may not
exploit the best possible delivery delay brought enter into high-power radio range and do not re-
by mobility. This scheme achieves the optimal duce the cost of other idle hardware components.
delay with unlimited relay buffers. However, Similar to (Jun et al., 2006), a throwboxs with
such a multiple-copy scheme generally incurs power management is proposed in (Banerjee et
significant overhead on storage constraint. Ip et. al., 2007) . In the architecture, the mobile node
al. (2007) proposed a buffer-management strategy, beacon their position, direction and speed using
RRFS-with-RandomDrop, to avoid head-of- line the long-range radio when it moves to the range
blocking in the FIFO case. They showed that the of tier-0, long-range radio. If the throwbox hears
proposed strategy can reduce the degradation of a beacon, it predicts how long the mobile node
average delivery delay performance. will contact with the tier-1, short-range radio. By
predicting the trajectory of an oncoming mobile
Device Energy node, the tier-0 wakes the tier-1 to undertake the
data transmission.
During opportunistic contacts, the nodes may need
to communicate in order to transfer the data to
destination. For conserving the device energy, the physical layer
node may need to know the contact time and the
duration for transfer the complete or segmented In order to operate under the challenged network
bundle in order to wake up at the contact time. Jun (Fall et al., 2003), the concept of long-lived is
et. al. (2005) proposed a power management in an important requirement. For some applica-
DTN depending on knowledge-based mechanisms tions, i.e. underwater sensor network, terrestrial
similar to the DTN routing algorithms (Jain et al., sensor network, they must design the suitable
2004). Two extremes of knowledge are complete communication hardware. Existing underwater
knowledge and zero knowledge. The former has communication systems have some useful physi-
the precise time of contacts, so it can provide the cal design for DTN. Freitag et al. (2005) have
upper bound on performance. The latter do not has developed a compact and low-power acoustic
any knowledge, so a node chooses a random time modem, called Micro-modem. The low-power
and broadcasts a beacon. If it does not receives design is an enabling factor for long-lived sensor
a beacon, the node sleeps and wakes up at the networks over battery-powered nodes. In (Wills
beginning of the next beacon time. In the partial et al., 2006), authors propose an inexpensive and
knowledge, a node knows statistics some informa- low-power hardware modem to save the power
tion about the contact durations and waiting times in order to provide long-lived operator. For low-
256
power operation, the power-saving is to use the The same process is repeated until the message
wakeup tone receiver to trigger the more expensive finally reaches its destination. This approach tries
data receiver. Authors also assume the radio will to reduce the buffer usages and the number of
be duty cycled, alternating on-and-off periods message transferred in the network. But it may
frequently when active, and with long-term off suffer the long delays and low delivery ratios.
periods when inactive. On the other hand, flooding-based approach may
generate multiple copies of the same message.
Each message can be routed independently for
mac layer increased efficiency and robustness. This ap-
proach achieves lower delays and higher delivery
As the section III, the medium access control ratio at the cost of a larger buffer space and more
(MAC) protocol needs to be modified for the high message transfers.
latency requirement. The communicating nodes
in DTN may suffer from high latency problem. Forwarding g-based approach
Similarly, the underwater acoustic communication
may also suffer from latencies larger than radio In the forwarding-based scheme, based on what
communication. In (Rodoplu et al., 2005), authors type of knowledge nodes use to select the appropri-
extend S-MAC’s schedule synchronization to ate or the best path to destination node, the prior
sender-receiver pairs in underwater. This protocol studies can be classified into three categories:
explains how to achieve a locally synchronized direct transmission, location-based, Knowledge-
schedule even in the presence of long propagation based and control-movement based.
delays. Each node schedules the time to transmit
the next packet, and broadcasts this information Direct Transmission
by attaching it to the current data packet. While
hearing the broadcasts, the other nodes will know Spyropoulos et. al. (2004) is proposed a simple
when to wake up for the subsequent packet. single-copy routing called direct transmission
However, in order to operate at a low collision routing. In this approach, after the source node
rate, each node requires a small duty cycle, which generates a message, the message is hold by
makes throughput low. the source node until it reaches the destination
node. The main advantage of this scheme is that
it incurs minimum data transfers for message
neTwork layer deliveries. On the other hand, although having
minimal overhead, this scheme may incur very
In this section, we discuss some routing solutions long delays for message delivery since the delivery
for a DTN. Based on the number of copies of a delay for this scheme is unbounded (Grossglauser
message forwarded by the node, we can define et al., 2002) .
two different routing schemes: forwarding-based
(single copy) approach and flooding-based Location-Based
(multiple copies) approach. In forwarding-based
approach, there is only one single custodian for In the location-based approach, nodes will choose
each message to help forwarding the message to the neighbors who are closest to the destination to
destination. When the current custodian forwards pass the message. LeBrun et al. (2005) proposed a
the copy to an appropriate next-hop neighbor, this method using the motion vector (MoVe) of mobile
neighbor becomes the message’s new custodian. nodes to predict their future location. The MoVe
257
scheme uses the knowledge of relative velocities (ED) where the contact oracle is available. The
of a node and its neighboring nodes to predict the contacts oracle contains information about con-
closest distance between two nodes. After the tacts between two nodes at any point in time.
nodes’ future locations are calculated, messages When this algorithm assumes the queuing delay
are passed to nodes that are moving closer to the is zero, the modified Dijkstra with time-varying
destination. As compared to epidemic routing, cost function based on waiting time is used to find
this approach has less control packet overhead the route. The third algorithm is Earliest Delivery
and buffer usage. Leguay et al. (2006) presented with Local Queuing (EDLQ), which uses the local
a strategy that uses a virtual coordinate routing queue occupancy to add an estimate of the queu-
called mobility pattern spaces (MobySpace). In ing delay to the ED algorithm. Final algorithm is
this approach, the node coordinates are composed the Earliest Delivery with All Queues (EDAQ)
of a set of dimensions, each dimension repre- where the contact oracle and queuing oracle is
sents the probability that a node will be found available. The queuing oracle gives information
in a specific location that is a virtual expression about instantaneous buffer occupancies (queuing)
of the mobility pattern and does not geographic at any node at any time. This algorithm adds the
coordinate of the node. There are various destina- queuing oracle to ED. This approach assumes that
tion functions are computed on this vector. They the accurate information about the oracle is known
showed that this approach consumes less resource in advance. This assumption may be workable
than epidemic routing when they deliver large for some scenarios where the node movement is
amount of bundles. predictable, i.e., city bus.
Musolesi et al. (2005) present the Context-
Knowledge-Based Aware Routing (CAR) protocol that provides an
asynchronous communication for message deliv-
In the knowledge-based approaches, based on ery. In a DTN, since the receiver is often not in the
certain knowledge about the network, the source same connected network, synchronous delivery
and intermediate nodes decide which nodes to of message is typically not possible. In CAR, if
forward messages as well as whether it should a message cannot be delivered synchronously,
transmit the message immediately or hold the the message is sent to a host that has the highest
message until it meets a better node. Jain et al. probability of successful delivery and acts as a
(2004) proposed the knowledge-based routing message carrier. The delivery probability process is
scheme which is the first study in this area. De- based on the evaluation and prediction of context
pending on the amount of knowledge about net- information using kalman filters. The prediction
work topology characteristics and traffic demand, process is used during temporary disconnection
they define four knowledge oracles. Each oracle and the process is continued until it is possible
presents some particular knowledge of network. to guarantee certain accuracy. In addition, the
Based on the available oracles, the authors pres- epidemic routing can be considered optimal in
ent a corresponding routing algorithm. The first terms of delivery ratio because each message is
algorithm is called Minimum Expected Delay propagated to all accessible nodes which have
(MED). In this algorithm, when the contact sum- large buffers to hold the messages. They showed
mary oracle, which contains information about in their simulations that if the buffer size is small,
aggregate statistics of the contacts, is available, the packet delivery ratio of CAR is batter than
Dijkstra with time-invariant edge costs based on the packet delivery ratio of epidemic routing due
average waiting time is used to find the best route. to that CAR only creates a single copy for each
The second algorithm is called Earliest Delivery message.
258
Burgess et al. (2006) proposed an effective the Message Ferrying approach is to introduce
routing protocol called MaxProp. A node uses non-randomness in the movement of nodes and
MaxProp to schedule packets transmission to its exploit such non-randomness to help deliver data.
peers and determines which packets should be Two variations of the MF schemes were devel-
deleted when buffer space is almost full. Packets oped. In the Node-Initiated MF (NIMF) scheme,
are scheduled based on the path likelihoods to ferries move around the deployed area according
peers according to historical data. In addition, to known specific routes and communicate with
several complementary mechanisms, including other nodes when they meet. With knowledge of
acknowledgments, a head-start for new packets, ferry routes, nodes periodically move close to
and lists of previous intermediaries are used in a ferry and communicate with the ferry. In the
this approach. They showed that their approach Ferry-Initiated MF (FIMF) scheme, ferries move
performs better than the protocols that have ac- proactively to meet nodes. It is assumed that the
cess to an oracle (Jain et al., 2004) that knows the ferry moves faster than nodes. When a node wants
schedule of meetings between peers. Kun et al. to send packets to other nodes or receive packets,
(Tan et al., 2003) proposed a shortest expected it generates a service request and transmits it to a
path routing (SEPR) similar to link-state routing chosen ferry using a long range radio. Upon recep-
to maintain a topology map to each other. SEPR tion of a service request, the ferry will adjust its
first estimates the link forwarding probability trajectory to meet up with the node and exchange
based on history data. When two nodes meet, they packets using short range radios. In both schemes,
exchange the link probability update messages nodes can communicate with distant nodes that
called effective path length (EPL). A smaller EPL are out of range by using ferries as relays. Zhao
value suggests a higher probability of delivery. et al. (2005) also propose multiple ferries with
When a node received a smaller EPL, it will stationary nodes to deliver data in networks and
update its local EPL value. EPL is also used in design of ferry routes. The route design problem
deciding which nodes to forward the messages. with multiple ferries is more complicated than the
Using SPER protocol, the same message could be single ferry case considering the possibility of in-
forwarded to multiple nodes to increase reliability teraction between ferries. The authors present four
and reduce delay. algorithms to generate ferry routes that meet the
traffic demand and minimize the weighted delay.
Control-Movement Based The authors considered algorithms that assume no
interaction between ferries, either using a single
In contrast to letting the mobile host wait pas- route (SIRA) or multiple routes (MURA). In the
sively for reconnection, the mobile hosts actively single ferry, they adapt solutions for the well-
modify their trajectories to minimize transmission studied traveling salesman problem (TSP). In the
delay of messages. Some works have proposed multiple ferries, the algorithm is to assign nodes
approaches that try to limit delay by controlling to specific ferries. The authors also considered
node mobility. Zhao et al. (2004) propose a Mes- algorithms that allow data relaying between ferries
sage Ferrying (MF) approach for data delivery in directly, ferry relaying algorithm (FRA), or indi-
sparse network. MF is a mobility-assisted approach rectly, node relaying algorithm (NRA). In NRA,
which utilized message ferries to provide commu- data is relayed between ferries via nodes, so the
nication service for nodes in the network. Similar NRA adopts a geographic approach for assigning
to their real life, message ferries move around the nodes to ferries. In FRA, data may be forwarded
deployment area and take responsibility for car- through multiple ferry routes while being routed
rying data between nodes. The main idea behind to the destination. But instead of relaying data via
259
nodes as in NRA, ferries exchange data directly. work. For this reason, Epidemic Routing incurs
Therefore, ferry routes need to be synchronized for significant demand on both bandwidth and buffer.
two ferries to meet each other. Simulation results To reduce such overhead, there are many related
are obtained to evaluate the performance of route paper to make epidemic routing consume fewer
assignment algorithms, especially on the effect resources (Spyropoulos et al., 2005 ; Lindgren et
of the number of ferries on the average message al., 2003 ; Tan et al., 2003). To bound the overhead
delay. Numerical results indicate that when the of delivering a message, Spyropoulos et al. (2005)
traffic load is low, the improvement in delay due proposed a technique called Spray and Wait to
to the increased number of ferries is modest. This control the level of flooding. In the spray phase,
is because the delay is dominated by the distance there are L numbers of copies that are initially
between nodes. However, when the traffic load spread over the network by the source node or
is high, an increase in the number of ferries can other nodes to L distinct relays. In the wait phase,
significantly reduce the delay. if the destination was not found during spray
phase, each node that has a copy of message will
Flooding-based approach perform direct transmission. Binary spray and wait
is a variation of Spray and Wait and produces a
In the flooding-based approach, each node broad- better performance. In this approach, the binary
casts the received packet to all of its neighbors, spray source node sends half of the copies of the
with the hope that one of these intermediate nodes message to the new relay node, and keeps the rest
will reach the destination. for itself. The source node and relay nodes repeat
this procedure until there is only one copy left.
Epidemic Routing When there is only one copy left, it switches to
direct transmission.
Epidemic routing is first proposed by Vahdat and
Becker (2000) for forwarding data in a DTN. Conditional Routing
Epidemic routing utilizes the epidemic algorithm
(Demers et al., 1987) that was originally proposed In conditional routing, nodes are not blindly
for synchronizing replicated databases. The epi- forward the messages to all or some neighbors.
demic algorithm ensures that a sufficient number Instead, nodes estimate the probability of each link
of random exchanges of data in the network and to destination and use this information to decide
guarantees all nodes will eventually receive all whether it should store the packet and wait for a
messages. Epidemic Routing works as follows. better chance as well as to decide which nodes
When two nodes come into contact, each node to forward. Lindgren et al. (2003) proposed a
will exchange the list of all message IDs that they probabilistic routing protocol, called PROPHET
have in their buffers, called the summary vector, to (Probabilistic Routing Protocol using History of
see if there are any messages that the other node Encounters and Transitivity). PROPHET estimates
has that it has not received. After such pair-wise probabilistic metric called delivery predictability.
exchange of messages, each node will get all the This metric indicates the probability of success-
messages carried by the other node that it has fully delivering a message to the destination from
not received. When this operation completes, the the local node. PROPHET operates in a similar
nodes have the same messages in their buffers. way as Epidemic Routing (Vahdat et al., 2000).
The Epidemic Routing is similar to the flooding When two nodes meet, they exchange summary
routing because it tries to send each message to vectors containing the delivery predictability vec-
all nodes using the summary vector in the net- tor which is based on the delivery predictability
260
Table 1. The comparison of routing protocols for DTN
Estimation of link Complexity of information exchange Reactive or

Protocol Buffer Management forwarding probability or computation for the link state Proactive
Epidemic Infinite No Don’t need Reactive
CAR Infinite YES, using Kilman filter Computation only Reactive
Spray and wait Infinite No Don’t need Reactive
YES, using delivery
PROPHET Infinite Exchange and computation Reactive
predictability vector
YES, estimating the deliv-
MaxProp Infinite Exchange and computation Reactive
ery likelihood
YES, using oracle based
Knowledge Infinite Exchange and computation Reactive
Dijkstra algorithm
YES, remove those packets
SEPR YES computation Reactive
with smaller EPL
Direction trans-
Infinite Don’t need Don’t need Reactive
mission
Exchange and computa-
MoVe Infinite Exchange and computation Reactive
tion
YES, message time out and
NIFM No Don’t need Proactive
buffer overflow
FIFN No No Compuation Proactive
information. In theory, if two nodes are often en- with extremely high latency such as deep space
countered, they have high delivery predictability communications. LTP handles the transmission of
to each other. On the other hand, if a pair of nodes a block of data, which can be split into segments
does not encounter each other in a while, they that match the maximum transmission unit for the
are intuitively not good forwarders of message link. In an Interplanetary Internet setting, LTP is
to each other. Hence, the delivery predictability intended to serve as a reliable “convergence layer”
values must age (i.e. be reduced) as time goes. protocol over single hop deep-space RF links.
They showed in their simulation results that the LTP implements ARQ of data transmissions by
communication overhead of PROPHET is lower soliciting selective-acknowledgment reception
than epidemic routing because PROPHET is only reports. When a client service transmits the seg-
sent to better nodes. ments of data, some are flagged as checkpoints.
When a checkpoint is received, the receiver
returns a report of cumulative reception for that
TransporT layer block. Reports acknowledge checkpoints and
either signal successful reception or else trigger
The existing transport layer protocols, such as retransmission. Farrell et al. (2005) proposed a
TCP are not suitable for an environment where generic transport protocol for DTN that use the
frequent disruption is a norm and end-to-end LTP extension mechanism (Farrell et al., 2007)
paths are typically not available. In (Ramadas to create an end-to-end capable transport protocol
et al., 2007), authors proposed the Licklider called LTP transport (LTP-T). The LTP extension
Transmission Protocol (LTP) that provides mechanism was originally defined to handle the
retransmission-based reliability over single links addition of authentication fields to LTP and al-
261
lows for the addition of both header and trailer bundle layer
extensions, up to a maximum of 16 (of each). In
this work, authors define a new set of extensions As mentioned at section II, the bundle layer is
of LTP about the transport protocol, i.e., source responsible for storing, carrying and forward-
address, destination address, estimated block size ing the data in DTN. Except from the unicast
and congestion notification etc. bundle delivery, multicast and anycast delivery
Since Bundle Protocol (Scott, K.et al., 2007) approaches are typically used when there are more
requires the services of a “convergence layer than one destination.
adapter (CLA)” which is an interface between
the common bundle protocol and a specific bundle delivery approach
internetwork protocol suite to send and receive
bundles using an underlying Internet protocol, In a DTN, applications utilize nodes to send or
then in (Demmer et al., 2006) the authors present receive data that is carried in bundles which can
one such convergence layer adapter that uses the be delivered to a group of nodes. When the group
well-known Transmission Control Protocol (TCP). size is greater than one, the delivery semantics
The TCP-based convergence layer (TCPCL) is may be either the anycast or multicast. For anycast
used to link two bundle nodes. The lifetime of delivery, a bundle is delivered to at least one and
a TCPCL connection will match the lifetime of preferable only one of the members in a group. On
its underlying TCP connection. In other words, the other hand, for multicast delivery, the bundle
a TCPCL connection is initiated when a bundle is intended to be delivered to all members in the
node initiates a TCP connection to be established same multicast group.
for the purposes of bundle communication. It is
terminated either when the TCP connection ends Anycast
due to one or both nodes actively terminating the
TCP connection, or when network errors causes In (Gong et al., 2006), authors defined an anycast
a failure of the TCP connection. In (Wood et al., semantics model and propose a routing metric,
2007), the authors showed that the TCP protocol called EMDDA (Expected Multi-Destination
does not make effective use of available link Delay for Anycast), for anycast. The semantics
capacity in a challenged environment like an op- models allow message senders to explicitly
portunistic network. In (Wood et al., 2007) the specify the intended receivers of a message. In
authors proposed use Saratoga (Wood et al., 2008) this study, the anycast routing algorithm is based
as convergence layer. Saratoga is a UDP based on the metric EMMA which accurately estimates
file transfer protocol that can be used to transfer the delay from a node to the nearest member of
bundles. To send a bundle, the local bundle agent the destined anycast group. EMDDA of a node
will either place bundles as files for Saratoga to to an anycast group is defined as the minimum
transfer from its directory that can be accessible value of Practical Expected Delays (PEDs), which
to both the bundle and Saratoga processes, or PED is the expected delay of taking different
otherwise use inter-process communication to paths, from the node to all the destination group
notify Saratoga of and provide a bundle to be members. When a message arrives at a node, but
transferred. the node is not an intended receiver of the anycast
message, the node will calculate its EMDDA to
the destination group.
262
Multicast unicast transfer to each of the estimated intended

receivers by encapsulating the original multicast
Due to the network partitions and opportunistic message. The source node also buffers the multi-
contacts, nodes are difficult to maintain a source- cast message and sends out new unicast messages
rooted multicast tree during the lifetime of a mul- when being informed of new intended receivers
ticast session. The traditional approaches may fail (2) In STBR (Static TreeBased Routing), mes-
to deliver a message when the link is unstable. In sages are forwarded along a tree in the graph that
(Zhao et al., 2005), the authors investigated four is rooted at the source and reaches all receivers.
approaches which are adopted from multicasting Due to that the route is static, the message needs
in the Internet or MANET. (1) UBR (Unicast- to wait for the next opportunity to be forwarded
Based Routing) uses unicast transfer to achieve the if a message misses a contact opportunity with a
multicast service. (2) In Broadcast-Based Routing node. This may cause significantly increase the
(BBR), messages will be flooded throughout the message delay. (3) In DTBR (Dynamic TreeBased
network in order to reach the intended receivers. Routing), nodes can determine the next-hop for-
(3) In Tree-Based Routing (TBR), messages are wards of a message dynamically based on current
forwarded along a tree in the DTN graph that is available information. (4) BBR (BroadcastBased
rooted at the source and reaches all receivers. Routing) always includes all nodes in the network,
Messages are duplicated only at nodes that have so messages are flooded throughout the network.
more than one outgoing path. (4) Group-Based (5) GBR (GroupBased Routing) uses the concept
Routing (GBR) uses the concept of forwarding of forwarding group for each message by comput-
group which is a set of nodes that are responsible ing a shortest path tree as in STBR. The group is
for forwarding the message. Messages will be a set of nodes that are responsible for forwarding
flooded within the forwarding group to increase the the message. Messages are forwarded by flooding
chance of delivery. Based on the above algorithms, within the forwarding group.
the authors present some algorithms for DTN are In (Ye et al., 2006), the authors proposed an
as follows. (1) UBR (Unicast-Based Routing) uses on-demand situation-aware multicast (OS-mul-
Table 2. The comparison of multicast protocols for DTN
Additional
Protocol Routing Type Tree Type Knowledge
technologies
UBR Unicast N/A N/A
STBR Tree, using shortest path Source-routed, static tree N/A
DTBR Tree, using shortest path Source-routed, dynamic tree N/A
BBR Broadcast N/A Complete knowl- N/A
edge or summary
Source-routed, static path to a of the link states
GBR Tree, using shortest path N/A
group
Source-routed, intermediate node
OS-multicast Tree, using shortest path dynamic rebuild the tree rooted N/A
itself
Estimate the High power trans-
Tree, using the contact prob-
CAMR Source-routed neighbor knowl- mission and message
ability estimate
edge ferry
Tree, using the highest delivery delivery predict-
EBMR Source-routed Directional antenna
predictability ability
263
ticast) approach. Initially, a source-rooted tree is in the boundary region to use directional antenna
constructed in the similar way as STBR. When a to find nodes in other regions if they cannot hear
node receives a bundle, it will dynamically rebuild any such node using omnidirectional antenna.
the tree rooted at itself to all the destinations based
on the current network conditions. Their simula- security
tion results show that OS-multicast can achieve
smaller delays and better message delivery ratios A draft DTNRG document presents the bundle
than DTBR due to that each node in DTBR only security protocol specification (Symington et al.,
forwards bundles to downstream nodes to reach 2007) and an additional draft document (Sym-
the receivers in its receiver list and OS-multicast ington et al., 2007) explains the rationale for the
always use all possible chances to forward the design choices made in the specification. The
bundles to all the destinations. specification describes three security blocks that
In (Yang et al., 2006), the authors proposed a can be added to bundles to provide different se-
context-aware multicast routing (CAMR) scheme curity services. The Bundle Authentication block
where nodes are allowed to use high power trans- (BAB) is used to provide authentication over a
missions when the node density (which is locally single hop by adding a message authentication
observed) drops below a certain threshold. Each code or a signature to the bundle. The Payload
node maintains contact probabilities using its Security block (PSB) is used to provide end-to-
2-hop neighbor information. This allows each end authentication in a similar fashion and the
node to deliver traffic without invoking a route Confidentiality block (CB) is used to encapsulate
discovery process if all receivers are within its encrypted payload of a bundle. Different combina-
two-hop neighbor. In addition, nodes are allowed tions of these three security headers can be used
to act as message ferries when they discover they simultaneously. Seth et al. (2005) propose the use
are in a very sparse neighborhood and then travel of hierarchical Identity Based Cryptography (IBC)
to closer to the next-hop for delivering bundles. (Boneh et al., 2003) to achieve end-to-end security.
The combined high-power route discovery process The authors observe that traditional PKIbased ap-
and message ferrying features allow CAMR to proach is not suited for disconnected networks,
achieve much higher multicast delivery ratio than since in DTN there do not have online access to
DTBR and OS-multicast schemes. an arbitrary receiver’s public key or certificate. In
In (Chuah et al., 2007), the authors build HIBC, different regions have sub-regions which
the multicast scheme on top of the PROPHET maintain their own PKGs. A user sent messages
(Lindgren et al., 2003), so this scheme is called with one PKG to a user of another PKG. The
encounter-based multicast routing (EBMR) messages are authenticated and protected using
scheme. This scheme has several enhancements to the trust relations between PKGs and standard
improve the delivery performance. First, each node techniques of HIBC. The identifier of a principal
selects as many nodes as needed with the highest can be based on existing well-known identifiers
delivery predictability to each of the multicast like e-mail addresses. However, in (Asokan et
receivers. If the next-hop can not be found, a node al., 2007), the authors argue that HIBC is not
will cache the data until the timer expires. When necessary because cellular operators already have
the timer expires, the node simply selects a node roaming agreements intended to enable cross-
with the highest delivery predictability to multicast domain operation.
receivers. The second enhancement allows nodes
264
applicaTion layer protocol is that the proxy has to execute the entire
SMTP protocol forwarding the information via
In a DTN, traditional applications fail to take the inter-proxy protocol. In (Hyyrylainen et al.,
advantage of the communication opportunities 2007), the authors describe an architecture to en-
offered by those opportunistic contacts. Hence, able mail communication in a heterogeneous en-
even the application is delay-tolerant in nature, vironment that combines traditional server-based
the overall application performance can still suffer mail delivery and opportunistic communications
significantly in a disconnection-prone environ- for different types of devices. In this architecture,
ment. We will discuss the application in DTN mail messages are sent in bundles into the DTN
from the common example, e-mail, to the more and carried toward a DTN mail gateway (DTN-
complex example, web service. The well-known MWG). The DTN-MWG is responsible to forward
paradigms on the Internet is E-mail because this and receive the mail between the infrastructure
application is delay-tolerant by large and e-mail network and the DTN network. The DTN-MWG
users are used to wait for hours or days for a re- and corresponding device could implement the
ply. However, given that the underlying transport Bundle Protocol to eliminate unnecessary process
protocol of e-mail (i.e. TCP) is not designed for a (Scott et al., 2005). In addition, three different
DTN, supporting e-mail in such an environment user equipment options can be configured on their
is still quite challenging. device. (1) To enable the user to continue using
Scott et al. (2005) proposed the use of DTN traditional email applications, the user laptop can
SMTP proxies to hide the disruptions between have a specialized email proxy to receive email
end users in a challenged network. This proxy from traditional and DTN connectivity options.
is responsible to help the client to perform its (2) User equipment may contain separate mail
work and exchanges the corresponding informa- application for both traditional mail and DTN
tion to a peer proxy. The peer proxy receives the mail. (3) User equipment may contain only the
information and sends it to the SMTP server. The DTN mail application.
drawback of this proxy-based approach for SMTP
Table 3. The comparison of application service for DTN
Protocol Application Additional entities Client Application Server

K. Scott E-mail DTN SMTP proxies Don’t modified Don’t modified
T. Hyyrylainen et al. E-mail Add DTN Mail Proxy in client
DTN mail gateway Don’t modified
device
Add another mail application for
DTN mail
Only receive the mail from DTN
mail gateway
K. Scott Web DTN-enable web proxy Don’t modified Don’t modified
J. Ott et al. Web Add the bundle pro-
Bundle router (BR) Add the bundle protocol
tocol
Add the bundle pro-
BR and Proxy Add the bundle protocol
tocol
BR and Gateway Don’t modified Don’t modified
A. Balasubramanian et al. Web Proxy Add the user query database Don’t modified
265
Supporting e-mail in a DTN system is straight- Furthermore, when the proxy awaits connection
forward since that fits into the characteristics of from a mobile node and has pending response, it
the DTN very well. Adding support for Web in downloads the responses and fetches some relevant
DTN is much more complicated, because highly web pages. In addition, the proxy will prioritize
interactive application protocols, such as HTTP, response bundles at next contact time.
are not well suited for this kind of environments. In
(Scott et al., 2005), Scott proposed an implement
of DTN-enable web proxy by extending the World conclusion
Wide Web Offline Explorer (WWWOFFLE,
2009). The authors split the WWWOFFLE proxy DTN is an emerging system that is getting grow-
and added client and a server side. The client side ing interest in networking research community.
links to the challenged network and uses DTN The DTN places different research challenges on
bundles to communicate with the server side. The different layers of a protocol stack. In this chapter,
server side has full connectivity to the Internet, we provide a quick overview of the state-of-the-
so that when the server receives requests from art work in providing solutions to various issues
clients, it can use HTTP to retrieve the requested in an opportunistic network.
web pages through the Internet. In (Ott et al.,
2006), authors presented a protocol design and a
system architecture for delay-tolerant access to reFerences
web pages. This work uses the bundle protocol
to transport the HTTP payloads in DTN network. Asokan, N., Kostiainen, K., Ginzboorg, P., Ott, J.,
Furthermore, several scenarios are proposed for & Luo, C. (2007). Applicability of identity-based
retrieving the content. First scenario is End-to-end cryptography for disruption-tolerant network-
HTTP-over-DTN. This scenario requires both cli- ing. In . Proceedings of MobiOpp, 07, 52–56.
ent and server implementing HTTP-over-DTN so doi:10.1145/1247694.1247705
that bundles can be sent directly to the respective Balasubramanian, A., Zhou, Y., Croft, B. W.,
server. Second scenario is Proxy-based HTTP- Levine, B. N., & Venkataramani, A. (2007). Web
over-DTN. Adding proxies into an HTTP-over- search from a bus. In Proceedings of the second
DTN may support a mobile node in content ag- workshop on Challenged networks CHANTS
gregation from one or more origin servers. Finally, CHANTS ’07, (pp. 59–66).
gatewaying HTTP-over-DTN proposes gateway
entities that communicate with another gateway Banerjee, N., Corner, M. D., & Levine, B. N.
through HTTP-over-DTN. The web clients and (2007). An energy-efficient architecture for dtn
web servers communicate with gateway through throwboxes. In . Proceedings of IEEE INFOCOM,
HTTP-over-TCP and the intermediary gateway is 2007, 776–784.
required to covert between HTTP-over-DTN and
Boneh, D., & Franklin, M. (2003). Identity based
HTTP-over-TCP operation. Balasubramanian et
encryption from the weil pairing. In SIAM Journal
al. (2007) proposed a system, called Thedu which
of Computing, 586–615.
use an Internet proxy to collect search engine result
and prefetch result pages. The mobile node can Brewer, E., Demmer, M., Du, B., Ho, M., Kam,
receive the user query through web interface and M., & Nedevschi, S. (2005). The case for tech-
store it until the mobile node contacts with the nology in developing regions. IEEE Computer,
proxy. If the connection is broken, the remaining 38, 25–38.
web pages will be downloaded at next contact time.
266
Burgess, J., Gallagher, B., Jensen, D., & Levine, Farrell, S., Ramadas, M., & Burleigh, S. (2007).
B. N. (2006). Maxprop: Routing for vehicle-based Licklider transmission protocol - extensions.
disruption-tolerant networks. In . Proceedings Internet-draft.
of IEEE INFOCOM, 2006, 1–11. doi:10.1109/
Freitag, L., Grund, M., Singh, S., Partan, J., Koski,
INFOCOM.2006.228
P., & Ball, K. (2005). The whoi micro-modem:
Burns, B., Brock, O., & Levine, B. N. (2005). Mv An acoustic communications and navigation
routing and capacity building in disruption toler- system for multiple platforms. In Proceedings of
ant networks. Proceedings of IEEE INFOCOM, the IEEE/MTS OCEANS Conference, Washington
1, 398–408. DC, (Vol. 2, pp. 1086–1092).
Chaintreau, A., Hui, P., Crowcroft, J., Diot, C., Gong, Y., Xiong, Y., Zhang, Q., Zhang, Z., Wang,
Gass, R., & Scott, J. (2005). Pocket switched net- W., & Xu, Z. (2006). Anycast routing in delay
works: Real-world mobility and its consequences tolerant networks. In IEEE GLOBECOM (pp.
for opportunistic forwarding. Technical report, 1–5).
University of Cambridge Computer Laboratory,
Grossglauser, M., & Tse, D. N. C. (2002). Mobility
Cambridge, UK.
increases the capacity of ad-hoc wireless networks.
Chen, Y., Zhao, W., Ammar, M., & Zegura, E. In IEEE/ACM Transactions on Networking, 10(4),
(2007). Hybrid routing in clustered dtns with mes- 477–486.
sage ferrying. In Proceedings of the 1st interna-
Hyyrylainen, T., Karkkainen, T., & Luo, C. (2007).
tional MobiSys workshop on Mobile opportunistic
Opportunistic email distribution and access in
networking MobiOpp ’07, (pp. 75–82).
challenged heterogeneous environments. In Pro-
Chuah, M., & Xi, Y. (2007). An encounter-based ceedings of the second workshop on Challenged
multicast scheme for disruption tolerant networks. networks CHANTS (pp.97–100).
Technical report published by the University of
Interplanetary internet project (n.d.). Interplan-
Lehigh, Lehigh, PA.
etary internet project. Retrieved from http://www.
Demers, A. A., Greene, D., Hauser, C., Irish, W., ipnsig.org
Larson, J., Shenker, S., et al. (1987). Epidemic
Ip, Y. K., Lau, W. C., & Yue, O. C. (2007). Forward-
algorithms for replicated database maintenance. In
ing and replication strategies for dtn with resource
Proceedings of the Sixth Symposium on Principles
constraints. In . Proceedings of IEEE Vehicular
of Distributed Computing, (pp. 1–12).
Technology Conference, 1, 1260–1264.
Demmer, M., & Ott, J. (2006). Delay tolerant
Jain, S., Fall, K., & Patra, R. (2004). Routing in
networking tcp convergence layer protocol.
a delay tolerant network. In Proceedings of ACM
Internet-draft.
SIGCOMM, (Vol. 34, pp. 145–158).
Fall, K. (2003). A delay-tolerant network archi-
Jun, H., Ammar, M. H., Corner, M. D., & Zegura,
tecture for challenged internets. In Proceedings
E. W. (2006). Hierarchical power management in
of ACM SIGCOMM, (pp. 27–34).
disruption tolerant networks with traffic-aware
Farrell, S., & Cahill, V. (2005). Ltp-t: A generic optimization. In Proc. ACM SIGCOMM Work-
delay tolerant transport protocol. Technical report, shop on Challenged Networks (CHANTS), (pp.
University of Dublin, Dublin. 245–252).
267
Jun, H., Ammar, M. H., & Zegura, E. W. (2005). Rodoplu, V., & Park, M. K. (2005). An energy-
Power management in delay tolerant networks: efficient mac protocol for underwater wireless
A framework and knowledge-based mechanisms. acoustic networks. In Proceedings of the MTS/
In IEEE Communications Society Conference on IEEE OCEANS’05 Conference.
Sensor and Ad Hoc Communications and Networks
Scott, K. (2005). Disruption tolerant networking
(SECON), (pp. 418–429, 2005).
proxies for on-the-move tactical networks. In IEEE
Krotkov, E., & Blitch, J. (1999). The defense ad- MILCOM 2005, (Vol. 5, pp. 3226–3231).
vanced research projects agency (darpa) tactical
Scott, K., & Burleigh, S. (2007). Bundle protocol
mobile robotics program. In The International
specification. Internet- Draft.
Journal of Robotics Research (pp. 769–776).
Segui, J., & Jennings, E. (2006). Delay tolerant net-
LeBrun, J., Chuah, C.-N., Ghosal, D., & Zhang,
working v bundle protocol simulation. In Second
M. (2005). Knowledgebased opportunistic for-
IEEE International Conference on Space Mission
warding in vehicular wireless ad hoc networks.
Challenges for Information Technology.
In IEEE Vehicular Technology Conference (VTC),
(pp. 2289– 2293). Seth, A., & Keshav, S. (2005). Practical security for
disconnected nodes. In Proceeding of Workshop
Leguay, J., Friedman, T., & Conan, V. (2006).
on Secure Network Protocols (pp. 31–36).
Evaluating mobility pattern space routing for
DTNs. In Proceedings of IEEE INFOCOM. Spyropoulos, T., Psounis, K., & Raghavendra, C.
S. (2004). Single-copy routing in intermittently
Lindgren, A., & Doria, A. (2007). Experiences
connected mobile networks. In 2004 First Annual
from deploying a reallife dtn system. In 2007 4th
IEEE Communications Society Conference on Sen-
IEEE Consumer Communications and Networking
sor and Ad Hoc Communications and Networks,
Conference, (pp. 217–221).
(pp. 235 – 244).
Lindgren, A., Doria, A., & Schelen, O. (2003).
Spyropoulos, T., Psounis, K., & Raghavendra,
Probabilistic routing in intermittently connected
C. S. (2005). Spray and wait: An efficient rout-
networks. ACM SIGMOBILE Mobile Comput-
ing scheme for intermittently connected mobile
ing and Communications Review, 7(3), 19–20.
networks. In Proceeding of the 2005 ACM SIG-
doi:10.1145/961268.961272
COMM workshop on Delay-tolerant networking
Musolesi, M., Hailes, S., & Mascolo, C. (2005). WDTN ’05, (pp. 252–259).
Adaptive routing for intermittently connected
Symington, S., Farrell, S., Weiss, H., & Lovell,
mobile ad hoc networks. In IEEE WoWMoM 2005
P. (2007). Bundle security protocol specification.
(pp. 183–189).
Internet-draft.
Ott, J., & Kutscher, D. (2006). Bundling the web:
Symington, S., Farrell, S., Weiss, H., & Lovell,
Http over dtn. In Proceeding of WNEPT2006.
P. (2007). Delay-tolerant networking security
Ramadas, M., Burleigh, S., & Farrell, S. (2007). overview. Internet-draft.
Licklider transmission protocol - specification.
Tan, K., Zhang, Q., & Zhu, W. (2003). Shortest path
Internet-draft.
routing in partially connected ad hoc networks.
In Proceedings of IEEE Globecom 2003, (Vol. 2,
pp. 1038–1042).
268
Tariq, M. M. B., Ammar, M., & Zegura, E. (2006). WWWOFFLE. (2009). Retrieved from http://
Message ferry route design for sparse ad hoc net- www.gedanken.demon.co.uk/wwwoffle/
works with mobile nodes. In Proceedings of the
Yang, P., & Chuah, M. C. (2006). Context-aware
seventh ACM international symposium on Mobile
multicast routing scheme for disruption tolerant
ad hoc networking and computing MobiHoc ’06
networks. In Proceedings of ACM international
(pp. 37–48).
workshop on PE-WASUN ’06 (pp. 66–73).
The zebranet wildlife tracker. (2009). Retrieved
Ye, Q., Cheng, L., Chuah, M. C., & Davison, B. D.
from http://www.princeton.edu/ mrm/zebranet.
(2006). Os-multicast: On-demand situation-aware
html
multicasting in disruption tolerant networks. In .
Tier. (2009). Retrieved from http://tier.cs.berkeley. Proceedings of IEEE VTC, 1, 96–100.
edu/wiki/Home
Zhang, Z., & Fei, Z. (2007). Route design for
Vahdat, V., & Becker, D. (2000). Epidemic routing multiple ferries in delay tolerant networks. In
for partially connected ad hoc networks. Technical Proceedings of IEEE Wireless Communications
Report CS-200006, Duke University. and Networking Conference (pp. 3460–3465).
Wills, J., Ye, W., & Heidemann, J. (2006). Low- Zhao, W., Ammar, M., & Zegura, E. (2004). A
power acoustic modem for dense underwater message ferrying approach for data delivery in
sensor networks. In Proceedings of the First ACM sparse mobile ad hoc networks. In Proceedings of
International Workshop on UnderWater Networks the 5th ACM international symposium on Mobile
(WUWNet), Los Angeles, CA, (pp. 79–85). New ad hoc networking and computing MobiHoc ’04
York: ACM. Retrieved from http://www.isi.edu/ (pp. 187–198).
ohnh/PAPERS/Wills06a.html
Zhao, W., Ammar, M., & Zegura, E. (2005).
Wood, L., Eddy, W. M., Ivancic, W., McKim, J., Controlling the mobility of multiple data transport
& Jackson, C. (2007). Saratoga: a delay-tolerant ferries in delay-tolerant network. In . Proceedings
networking convergence layer with efficient link of IEEE INFOCOM, 2, 1407–1418.
utilization. In Proceeding of IWSSC’07.
Zhao, W., Ammar, M., & Zegura, E. (2005).
Wood, L., McKim, J., Eddy, W., Ivancic, W., & Multicasting in delay tolerant networks: Semantic
Jackson, C. (2008). Saratoga: a convergence layer models and routing algorithms. In Proceeding of
for delay-tolerant networking. Internet draft. Sigcomm Workshop in DTN, (pp. 268–275).
Wood, L., Peoples, C., Parr, G., Scotney, B., &
Moore, A. (2007). Tcps protocol radius:the dis-
tance where timers prevent communication. In
Proceeding of IWSSC’07.
269
Section 6
271
Chapter 16
Simple Transportation
Management Framework
Chyi-Ren Dow
absTracT
The Simple Transportation Management Framework (STMF) specifies a set of rules and protocols which
can be used to organize, describe, and exchange transportation management information between trans-
portation management applications and equipments. The STMF framework consists of four elements,
including Management Information Base (MIB), Structure and Identification of Management Information
(SMI), Simple Network Management Protocol (SNMP), and Simple Transportation Management Protocol
(STMP). MIB is a collection of management objects written in ASN.1 notation. SMI is the definition of
how to create management objects and a hierarchical definition of nodes where management objects will
be attached for unique identification. SNMP is a communications protocol for configuring and monitor-
ing of network devices. STMP is a variation of SNMP to address low-bandwidth communication links
and real-time device monitoring.
inTroducTion inter-operability, and maintenance of STMF. The

STMF framework consists of the following four
The STMF specifies a set of rules and protocols elements. The brief definitions of these elements
which can be used to organize, describe and ex- are listed below and the detailed descriptions and
change transportation management information relationships of these elements will be given in the
between transportation management applications following sections.
and equipments. STMP is based on the Internet-
standard Network Management Framework and • Management information base (MIB)
its purpose is to provide a high-level compatibility, (Perkins et al., 1997): A collection of man-
agement objects written in ASN.1 (Abstract
DOI: 10.4018/978-1-60566-840-6.ch016 Syntax Notation One) (Steedman et al.,
Simple Transportation Management Framework
Figure 1. Relationships of SMI, MIB, and ASN.1
1993; ITU-T X.680-X.690, 1994) notation, definitions (ASN.1 specifications) are included in
which is a standard and flexible notation a MIB module. As shown in Figure 1, the differ-
that describes data structures for represent- ence is made because SMI defines how to create
ing, encoding, transmitting, and decoding managed objects and how to utilize ASN.1 in order
data. to create and identify management information
• Structure and identification of manage- (MIB objects) within a tree-like structure. Man-
ment information (SMI): The definition agement center uses Structure and Identification
of how to create management objects and of Management Information to define MIB, then
a hierarchical definition of nodes where using BER or OER encoding scheme to generate
management objects will be attached for SNMP or STMP.
unique identification.
• Simple network management protocol smi
(SNMP) (Stallings et al., 1993; Stallings
et al., 1996; Feit et al., 1995): A commu- Managed objects would be accessed via MIB and
nications protocol developed by the IETF objects in the MIB would be defined using ASN.1
for configuring and monitoring of network which should be in conformance with IAB STD
devices. 16 (RFC 1212). Each object type would have a
• Simple transportation management pro- name, syntax, and an encoding. The OBJECT
tocol (STMP): A variation of SNMP devel- IDENTIFIER would represent a unique name. An
oped by NEMA to address low-bandwidth OBJECT IDENTIFIER should be administratively
communication links and real-time device assigned a name. The administrative policies dis-
monitoring. NEMA is the trade association cussed in RFC 1212 would be used for assigning
of choice for the electrical manufacturing names and identifiers. When transmitted on the
industry, and it provides a forum for the de- network, the encoding of an object type determines
velopment of technical standards how its instances are represented.
relationships of smi, Names

mib, and asn.1
Names are used to identify managed objects.
SMI describes the common structures and identi- This sub-clause specifies names that should be
fication schemes for the definition of management hierarchical in nature. The OBJECT IDENTIFIER
information. ASN.1 is used to specify SMI and it concept is used to model this notion. OBJECT
can be compiled by MIB compilers. To do so, SMI IDENTIFIERs can be used to identify objects,
272
Figure 2. nema of OBJECT IDENTIFIER

In this example, four nodes are assigned under
the nema node. The definitions are listed below.
• nemaMgmt OBJECT IDENTIFIER::=

{nema 1}
• nemaExperimental OBJECT
IDENTIFIER::= {nema 2}
• nemaPrivate OBJECT IDENTIFIER::=
{nema 3}
• transportation OBJECT IDENTIFIER::=
{nema 4}
SYNTAX
The SYNTAX is used to define the structure and

it is corresponding to object types. RFC 1155 can
be used to define this structure. The SYNTAX
type consists of primitive types and guidelines
for enumerated INTEGERs and constructor
types.16.1.2.3 Managed Objects
IAB STD 16 specifies a format to be used by
documents that defines objects. All objects would
regardless of the semantics associated with the be defined using the OBJECT-TYPE macro de-
object (e.g., a network object, a standards docu- scribed in RFC 1212.
ment, etc.). The OBJECT-TYPE macro definition contains
An OBJECT IDENTIFIER is a sequence of at least five fields, including object name, syntax,
integers which can be used to traverse a global access, status and description. Figure 3 shows an
tree. The tree consists of a root connected to a example of managed object. The “Access” can be
number of labeled nodes via edge and each node read-only, read-write, write-only or no-accessible.
may have branches of its own that are labeled. The “Status” can be mandatory, optional, obsolete,
This process may continue to an arbitrary level or deprecated.
of depth. A label is a pairing of a brief textual
description and an integer. The format used to mib
show this pairing shall be text (integer).
The IAB has transferred authority for the nema The MIB version 2, published in RFC 1213 in
node to NEMA. As shown in Figure 2, the nema 1991, steped up a number of useful variables
node is defined below using the ASN.1 syntax. missing from MIB-I. It has four characteristics: the
first one is that has no relationship with network
nema OBJECT IDENTIFIER::= management protocol (ITS Standards et al., 2003;
Aidarous et al., 1994). The second is each object
{ iso(1) org(3) dod(6) internet(1) private(4) has a unique Object Identifier (OID). The third
enterprises(1) 1206 } is that defined by SMI. The final one records the
type and access limits of authority.
273
Figure 3. ASN.1 Macro
As shown in Figure 4, the Object Naming Transportation management protocol

Tree is used to record all objects in MIB-II. If we
choose tcpConnState for instance, the location Figure 5 shows the comparison of SNMP, STMP
of tcpConnState is iso(1)→ org(3)→ dod(6)→ and SFMP. The STMP is based on SNMP. Several
internet(1)→ mang(2)→ mib-2(1)→ tcp(6)→ tcp- features are modified so that STMP can provide
ConnTable(4)→ tcpConnState(1), so we can know more efficient operations. STMP can be used to
that the OID of tcpConnState is 1.3.6.2.1.6.4.1. save bandwidth and offer all functions of SNMP
From Figure 4, we can also see some groups in under the infrequent information demands. Thus,
MIB-II, including system(1), interface(2), at(3), STMP is the subset of SNMP and, the STMP man-
ip(4), icmp(5), tcp(6), udp(7), egp(8), cmot(9) aged system can also use SNMP to carry on the
and transmission(10). communications. The major advantage of STMP
is that it uses an effective code method that is sup-
Figure 4. Object naming tree
274
Figure 5. Comparisons of SNMP, STMP and SFMP
ported by dynamic objects to reduce the overhead management systems to monitor network-attached
of package. The user can also set any informa- devices for conditions that warrant administra-
tion object by herself/himself through dynamic tive attention. It consists of a set of standards for
objects. Thus, STMP is the choice which is more network management such as application layer
elasticity and saves more bandwidth. However, protocol, a database schema, and a set of MIB.
compared with SNMP, it is much more difficult
in practice for STMP. snmp architectural model
TMP allows three protocols to coexist while
using the same protocol identifier. This was The SNMP architectural model is a collection of
achieved as a result of the fact that all SNMP network management stations and network ele-
messages start with an initial byte of 0x30 (e.g., ments. It consists of three key elements, including
SNMP uses Basic Encoding Rules and all SNMP Management stations, multiple network elements,
messages are defined as a SEQUENCE of data). and Network Management Protocol.
TMP has been designed to use the value of this A management station can perform manage-
first byte to identify which protocol is being ment applications which monitor and control
referenced. 0x30 is a value which identifies an network elements. Multiple network elements
SNMP message. Both SFMP and STMP messages are devices such as hosts, routers, and gateways
use the highest order four bits of the first byte to that contain agents to perform the network man-
determine the type of message (e.g., get request agement functions requested by the network
and set request). The lowest order four bits are management stations. The last, Simple Network
used to identify whether the message is a fixed Management Protocol is used to exchange network
message or one of the 13 dynamic objects. The management information between the network
table in Figure 6 defines the specific mapping of management stations and the agents.
the first byte value. The SNMP architectural model is shown in
Figure 7. The upper parts of the management
station and network elements are manager and
snmp agents, respectively. The lower part is responsible
for the information exchange between manage-
IETF defined SNMP which is a part of the protocol ment station and the different network elements.
suite to describe the managed objects contained Each component of SNMP is described in the
in MIB. This simple protocol is used in network next section.
275
Figure 6. Mapping of the first byte value
management components • Manager: Manager station that runs man-

agement applications to provide an inter-
There are four management components in SNMP, face. The network manager can monitor
including Manager, Agent, Network Protocol and and control the network by the interface.
Management Information Base. The components • Agent: The software that responds to re-
are described as follows: quest for information from the manager
station and according to the request from
manager station to act. Therefore, network
276
Figure 7. SNMP Architectural Model
Figure 8. MIB architectural diagram

device should be equipped with agent soft-
ware so that they can be managed from a
manager station.
• Network management protocol: The
communication protocol between manager
stations and agents. It provides a standard
way to exchange information between
manager stations and agents.
• Management information base: A col-
lection of Manager Objects (MO) that are
managed in remote devices. As shown in
Figure 8, each MO is a data variable and
each agent is a network that maintains an quest, GetResponse, and Trap. Figure 9 shows the
MIB. A manager action is either “monitor” relationship between services and management
or “control”. To monitor resources in a re- elements. Manager sends commands such as Get,
mote device, we can read the value of MOs Set, GetRequest, SetRequest and GetNextRequest
in the MIB. In the MIB, we can modify the to agents and then agent responds GetResponse.
values of MOs to “control” resources in a If agent occurs error during runtime, agent ap-
remote device. plication submits Trap to manager.
There are two versions of SNMP, including
SNMP v1 and SNMP v2. SNMP v2 provides more
snmp service services than SNMP v1.
SNMP provides four services and five SNMP SNMP v1

Protocol Data Units (PDUs). The four services
are Get, Set, GetNext and Trap. The five SNMP As shown in Figure 10, four services provided by
PDUs are GetRequest, SetRequest, GetNextRe- SNMP v1 are Get, Set, GetNext and Trap. Below
277
Figure 9. Services of SNMP
Figure 10. SNMP v1 services
are detailed descriptions about the five SNMP SNMP v2

PDUs: GetRequest, SetRequest, GetNextRequest,
GetResponse, and Trap. There are two new services provided by SNMP v2:
GetBulk and Inform. Centers use GetBulk com-
• Get Request: Used to retrieve the values mand to retrieve large amount of MIB objects data
of objects in the MIB of an agent. while different devices use Inform command to
• Get-Next Request: Used to retrieve the provide information to the management center.
values of the next objects in the MIB of
an agent. The management station can use snmp packet
this command through all MIB tree paths
to retrieve all variables. As shown in Figure 11, the SNMP packet consists
• Get Response: When an agent got Get of three fields: Version, Community Name and
Request from manager, the agent responds Protocol Data Unit.
the value to manager.
• Set Request: Used to set or modify the • Version: The number of SNMP version
values in the MIB of an agent. that is used by Manger and Agent. The “0”
• Trap: Used to report extraordinary events is used to indicate SNMP v1. The “1” is
to the manager if an error occurs on the re- used to indicate SNMP v2.
mote device. • Community name: It is the password of
SNMP. The Manager and Agent should
Figure 11. SNMP packet
278
Figure 12. PDU for GetRequest, GetNextRequest and SetRequest
use the same Community Name or else the Error Index indicates that the index of Var-
frame will be discarded. BindList is wrong. The type of Error Index is inte-
• Protocol data unit (PDU): It consists of ger. For the request from management, the values
some data that would be used in commu- of Error Status and Error Index will be 0.
nication such as Request ID, Error Status VarBindList is a list that combines Variable
and VarBindList. IDs and Variable Values. The architecture of
VarBindList is shown in Figure 13. The variable
PDU ID field is an Object ID and the field of Variable
Value may be an integer, octstring or IP Address.
As shown in Figure 12, the PDU for Get and Set The Object ID and the type of Variable Value are
service has four fields, including Request ID, Er- defined in MIB as mentioned in MIB earlier in
ror Status, Error Index and VarBindList and these the chapter.
fields are described in detail as follows:
SNMP Trap PDU
• RequestID: An integer type is used as an
identification. It identifies different ser- The Trap command is submitted from agent. The
vice actions in communication for SNMP architecture of Trap PDU is shown in Figure 14.
operation. The Trap PDU consists of six fields, including
• Error Status: Used to confirm whether this Enterprise, Agent Address, Generic Trap Num-
PDU is correct. The value will be equal to ber, Specific Trap Number, Time Stamp and
zero if this is a “Get-Request” command. VarBindList. The detailed descriptions of the six
If this command is “Get-Response”, the fields are as follows:
value of Error status will have six possible
values as shown in Table 1. • Enterprise: Type of object generating
trap.
• Agent address: Address of object generat-
ing trap.
• Generic trap number: Using integer to
Table 1. Values of error status and its meaning represent seven Trap types defined in RFC
1157.
Value of Error Status Meaning
• Specific trap number: The specific Trap
0 No Error
type defined by enterprise.
1 PDU has too many bytes
• Time stamp: Time elapsed between the
2 There is no object with this name
last initialization of the network entity and
3 Identifying the PDU type is bad the generation of the trap.
4 Incorrect implementation of SNMP
5 Unspecified errors of other types
279
Figure 13. VarBindList

Set-Request Command
Set command writes MIB object value. For ex-

ample, if we have to change the system description
(sysDescr), the command can be set as below:
Set-Request (sysDescr = name)

• VarBindList: Specific or all informa-
tion that can be used to solve this trap Set-Request command in SNMP can be used to
problem. add or delete rows of table. The result of the Set-
Request is related to the real implementation.
Get-Request Command
SNMP Example
SNMP uses index object to select column. The
index object of ipRouteTable is the ipRouteDes- Figure 15 shows a SNMP example of a traffic
tination and the entity of ipRouteNextHop is light controller. The center gets the remaining
1.3.6.1.2.1.4.21.1.7.10.3.4.5. time from the traffic light controller. MIB can only
While using SNMP command interface, the send a request each time. While the traffic center
Get-Request command should be inserted as requests information from the controller, the traf-
below: fic controller will request a DeviceNum from the
center. Then, the center will send TimePhaseNum
• Get-Request(1.3.6.1.2.1.4.21.1.7.10.3.4.5) to the controller and wait for response.
or The traffic center can request MaxGreen-
• Get-Request(ipRouteNextHop.10.3.4.5) Time, MinGreenTime, YellowTime, GreenTime,
RedTime from the traffic light controller. Each
Get-Next-Request Command request will get a response from the controller in
the same way.
GetRequest command is used to request one ob-
ject of a device per time. Thus, to request all of
the devices, Get-Next_Request can be used. The sTmp
command can be inserted as below:
This section introduces the Simple Transporta-
• Get-Next-Request tion Management Protocol and the comparison
(1.3.6.1.2.1.4.21.1.7.10.3.4.5)or between Simple Transportation Management
• Get-Next-Request Protocol and Simple Network Management
(ipRouteNextHop.10.3.4.5) Protocol. STMP is based on SNMP and supports
dynamic object that is a group of data elements.
Figure 14. SNMP trap PDU
280
Figure 15 Example of a traffic light controller
Besides, it also supports PMPP defined in the sub Dynamic Object Configuration Table
network of NTCIP.
Dynamic object configuration table consists of
dynamic object three elements: dynObjNumber, dynObjCon-
figOwner, and dynObjConfigStatus.
Dynamic object is a simple grouping of data ele- The dynObjNumber provides a number for
ments. It consists of a set of related MIB objects. the frequently used dynamic objects. The main
For example, object name, object phase, and purpose of the dynObjNumber is to identify which
object time value are dynamic objects of the time of the 13 dynamic objects this row of the table
controller of a traffic light. These three objects are is associated.
related to each other. Each dynamic object has an owner. There-
Between management station and the agent, fore, dynObjConfigOwner is used to indicate the
dynamic object increases control flexibility. A identity of the owner that defined the dynamic
dynamic object can combine all related objects object.
into a set. Thus, different tasks can be done easier Dynamic object records status of each managed
using a dynamic object. objects. Thus, dynObjStatus is used to indicate
To reduce the need of bandwidth during the the status of the dynamic object.
transformation, dynamic object is defined at run Table 2 depicts the composite table for dynamic
time. It is an effective way for the management object configuration and definition. The table in-
station to communicate with different agents. cludes the three dynamic object (dynObjNumber,
However, dynamic object can only access using dynObjConfigOwner, and dynObjConfigStatus),
the STMP protocol. dynObjIndex, and dynObjVariable.
281
Table 2. Composition table for dynamic object
dynObjNumber dynObjConfigOwner dynObjConfigStatus dynObjIndex dynObjVariable

1 <Owner of Dynamic Object #1> <OID of 1st object in
1
dynObj 1>
<OID of 2nd object in
2
dynObj 1>
<Status of Dynamic Object
<OID of 3rd object in
#1> 3
dynObj 1>
… …
<OID of 255th object in
255
dynObj 1>
1
dynObj 2>
2
dynObj 2>
#2> 3
dynObj 2>
… …
255
dynObj 2>
1
dynObj 3>
2
dynObj 3>
#3> 3
dynObj 3>
… …
255
dynObj 3>
… … … … …
1
dynObj 13>
2
dynObj 13>
#13> 3
dynObj 13>
… …
<OID of 255th obj -
255
dynObj 13>
The dynamic object is the combination of Objects in the set have already defined in the ISO
managed objects. Each dynamic object is given naming tree. Thus, dynObjVariable records each
an index to represent the sequence of the objects. managed object’s OID, which is the location of
The dynObjIndex column indicates the index of the managed object in the ISO naming tree.
the dynamic objects.
The dynamic object method treats all frequently
used related objects as a set of dynamic objects.
282
Dynamic Objects and System Operation Set Request-No Reply is similar to Set Request,
but the device will not return any response after
STMP supports 13 dynamic objects for each the request is sent.
agent. With a different set of dynamic objects, the Get Response and Set Response are used to
management station could configure each device. response to get and set requests command from
In practice, most management stations are likely centers. Trap Response is to inform the centers
to configure similar devices with similar dynamic that some situations occurred on a device.
object definitions. Get Error Response will response to the
centers that get request command is failed. Set
State Transition for dynObjConfigStatus Error Response will response to the centers that
set request command is failed.
Table 3 shows the state transition for dynObjCon-
figStatus. If no action takes place and response STMP PDU Field
indicates no Error. When the state changes to Figure 16 shows the STMP PDU fields. There are
invalid, all entries associated with the Confi- four fields in the STMP PDU: PDU format, mes-
gEntryStatus object are deleted or cleared and sage type, object ID and information field.
response indicates no error. The PDU Format is 0 or 1 (bit 7). If PDU
If no action takes place but response indicates format is “1”, it represents the STMP service. If
badValue. If Dynamic Object Validation succeeds the data is “0”, it represents a SNMP service or
then state changes to valid and response indicates it might be reserved for future use.
no error. If Dynamic Object Validation fails then Message type and object ID are header fields.
state remains underCreation and response indi- Message type is represented by bits 6-4. It indi-
cates genErr. The state changes to underCreation cates the STMP service. Below are functions of
and the response indicates no error. different message types:
STMP Service • 000: An STMP-GetRequest-PDU is con-

tained in the packet.
Services provided by STMP are Get Request, Set • 001: An STMP-SetRequest-PDU is con-
Request, and Set Request-No Reply. The service tained in the packet.
responses are Get Response, Set Response, Trap • 010: An STMP-SetRequest-NoReply-PDU
Response, Get Error Response, and Set Error is contained in the packet.
Response. • 011: An STMP-GetNextRequest-PDU is
Get Request is to get an object’s information. contained in the packet.
Set Request is used to set an objects information.
Table 3. State Transition for dynObjConfigStatus
Requested State
Invalid UnderCreation Valid
Invalid Invalid (1) UnderCreation (2) Invalid (3)
Valid (2) or
Current State UnderCreation Invalid (2) UnderCreation (3)
underCreation
Valid Invalid (2) Valid (3) Valid (1)
283
Figure 16. PDU field
• 100: An STMP-GetResponse-PDU is con- • 0001-1101: ID of STMP 13 dynamic

tained in the packet. objects
• 101: An STMP-SetResponse-PDU is con- • 1110: Reserved by TMP for future use.
tained in the packet. • 1111: Reserved by TMP for future use.
• 110: An STMP-ErrorResponse-PDU is
contained in the packet. The information field contains information that
• 111: Reserved by TMP for future use. has to send between the server and the agent. It
would be empty for STMP-GetRequest, STMP-
The object ID carries the ID of the object. It GetNextRequest, and STMP-SetResponse. The
is represented by bits 3-0. Below are descriptions PDU Information field for STMP-GetResponse,
about the data of object ID: STMP-SetRequest, and STMPSetRequestNoRe-
ply would consist of a series of component fields,
• 0000: Reserved by TMP for SFMP. each encoding one Referenced Object. Besides,
Figure 17.Configuring time dynamic object
284
the component fields would be encoded in order. object as variable 3.2 and so on. This can
It is according to the associated dynObjIndex, be shown in Figure 17.
with the first field encoding the value of the first • Finally, set this dynamic object to “valid”.
Referenced Object of the Dynamic Object, and the Now we can use this dynamic object to
last field encoding the value of the last Referenced transfer data.
Object of the dynamic object.
Examples Example: Getting a Dynamic Object

The examples in the following will provide more
detailed explanation about the STMP and the After finishing the setting of dynamic object, we
dynamic objects. can use this dynamic object to transfer data with
services such as “request” service. To continue
Example: Configuration above example, we are going to discuss the details
Time Dynamic Object of STMP PDU with request command. As shown in
Figure 18, we can see the actual content of STMP
Before using STMP to transfer object, we should “GetRequest” and “GetResponse” PDU.
set Dynamic Object Configuration Table. For In the Information fields, there are four vari-
example, we configure the third object as Time ables that we set in above example. They are
object. As shown in Figure 17, we should set the presented as the following:
third object is “invalid” first, and then we set it as
“underCreation”. Now we just can set the property • 3A 24 63 20 variable 1 = globalTime.0 =
of dynamic object as the following: November 29, 2000 at 2:00 am UTC
• 03 variable 2 = globalDaylightSaving.0 =
• Set the owner of this dynamic object to 3 = enableUSDST
“FCU”. • FF FF B9 B0 variable 3 = controller-stan-
• Set the variables of the dynamic object with dardTimeZone.0 = -18000 = EST
MIB object. In this example, the dynamic • 06 53 61 6D 70 6C 65 variable 4 = event-
object includes “GlobalTime” object as ClassDescription.1 (6 bytes) = “Sample”
variable 3.1 and “GlobalDayLightSaving”
Figure 18. Actual content of STMP PDU
285
Application Level: STMP be set after the dynObjConfigStatus is set as

“underCreation”.
Figure 19 shows an example of configuring traffic The traffic center starts setting the dynamic
light controller of dynamic object. First, the traffic objects (e.g., dynObjConfigOwner and dynObj-
center sends a request to the traffic light controller. Variable) when the traffic light controller sends
The request demands the traffic light controller a response. After all dynamic objects are set; the
to set the dynObjConfigStatus as “invalid”. The traffic center has to set the dynObjConfigStatus
center sends another request to the controller to as “valid”. Finally, setting the dynamic objects
set the dynObjConfigStatus as “underCreation” has done.
when the traffic center gets response from the After the setting of dynamic object is finished,
traffic light controller. The dynamic object can we can use it to actual applications. An applica-
Figure 19. Example of configuring traffic light controller of dynamic object
Figure 20. Application for traffic light using dynamic object
286
Figure 21. Comparisons between SNMP and

has Version number and Community Name as a
STMP
password.
Because SNMP uses more bandwidth than
STMP. SNMP can be used in the environment
that has weak probability of controlling object
and high network bandwidth. In comparison with
SNMP, STMP uses dynamic objects to reduce the
data transfer and efficiently use bandwidth.
Therefore, STMP can be used in the envi-
ronment and it has low network bandwidth and
frequent control of object value.
In summary, SNMP is simple and easy to imple-
tion for traffic light is shown in Figure 20. This ment. STMP uses Dynamic Composite Objects so
application uses dynamic object to control traffic that STMP uses bandwidth efficiently.
light. Figure 20 shows the traffic center submits a
get request to the traffic light controller and then aTms system for sTmp
the traffic light controller returns the value of
traffic light. In Figure 20, the “dynObjVariable” The Advanced Traffic Management System
word is the object number of dynamic object. The (ATMS) (National Highway Institute, 2001) is
traffic center uses getRequest service with object based on an open architecture. It uses the latest ITS
number to request the value of traffic light; the technologies to provide real-time transportation
traffic light controller would then responses to the system control, monitoring, and information.
traffic center with “getResponse” command. ATMS can detect the traffic situation and
As shown in Figure 21, the difference be- transport to traffic control center by the Internet
tween SNMP and STMP package is that SNMP (Rose et al., 1990). It can collect and manage
Figure 22. Operation process between message set and data dictionary
287
Table 4.
ParameterType MessageType
Length Search Activily

ObjectName ObjectID Type (byte) Range Config Search Response Response
12H 2BH 14H
globalDeviceSysTime globalDeviceSysTime 0F+02
0F+12 0F+42 0F+C2
ROCYear DeviceROCYear Integer 1 0~99
Month deviceMonth.0 Integer 1 1~12
Day deviceDay.0 Integer 1 1~31
Wekday deviceWeekday.0 Integer 1 1~7
Hur deviceHour.0 Integer 1 0~23
Minute deviceMinute.0 Integer 1 0~60
Second deviceSecond.0 Integer 1 0~60
DeviceState globalDeviceState 0F+41 0F+C1
10H
HWReset deviceHWReset.0 Constant 1 1
0F+10
11h
CommReset deviceCommReset.0 Constant 1 1
0F+11
2CH 2CH 2EH
LoopTestParameter deviceLoopTest.0 Integer 32 0~255
0F+47 0F+47 0F+C7
21H
DbLockState deviceDbLockState.0 Integer 1 0~2 0F+46 0F+C6
0F+16
18H 29H
EquipmentId deviceEquipmentId.0 Integer 2 0~ffh
0F+40 0F+C0
1FH
Password devicePassword.0 Char 6 ASCII 0F+45 0F+C5
0F+15
DeviceEvent globalDeviceEvent
1EH
RestartEvent deviceRestartEvent
0F+00
RestartMonth deviceRestartMonth.0 Integer 1 1~12
RestartDay deviceRestatDay.0 Integer 1~31
traffic situation to improve traffic efficiency and Reversible Lane Management, and Road Weather
safety. Information System.
The major services of ATMS consist of Net-
work Surveillance, Probe Surveillance, Surface ATMS Data Dictionary
Street Control, Freeway Control, HOV Lane
Management, Traffic Information Dissemination, Figure 22 shows the operation process between
Regional Traffic Control, Incident Management message set (IEEE Std 1488- 1999, 2000) and
System, Traffic Forecast and Demand Manage- data dictionary (IEEE Std 1489-1999, 1999). It
ment, Electronic Toll Collection, Emissions is based on Open Systems Interconnection (OSI)
Monitoring and Management, Virtual TMC and communication protocol and supplies the common
Smart Probe Data, Parking Facility Management, data format defined between applications. It can
ensure the communication of systems. The OSI
288
Reference Model is an abstract description for Perkins, D., & McGinnis, E. (1997). Understand-
layered communications and computer network ing SNMP MIBS. Upper Saddle River, NJ: Prentice
protocol design. It was developed as part of the Hall, Inc.
OSI initiative. As shown below, in its most basic
Rose, M. T. (1990). The Open Book: A Practical
form, it divides network architecture into seven
Perspective on OSI. Upper Saddle River, NJ:
layers which, from top to bottom, are the Applica-
Prentice Hall, Inc.
tion, Presentation, Session, Transport, Network,
Data-Link, and Physical Layers. Stallings, W. (1993). SNMP, SNMPv2 and CMIP
The Practical Guide to Network Management
ATMS C2F Level Definition Standards. Reading, MA: Addison-Wesley Pub-
lishing Company, Inc.
C2F uses the communication protocol standard
Stallings, W. (1996). SNMP, SNMPv2 and RMON.
of SNMP and STMP. The messages must be de-
Reading, MA: Addision-Wesley Publishing
fined as data objects and dynamic objects based
Company, Inc.
on ASN.1.
NTCIP Standard 1103 (2005). TS 3.2-1996 NTCIP
Example of Global Object Definition Simple Transportation Management Framework
- Amendment 1.
Using global object can setup time, database status
ITS Standards Outreach, Education and Training
on devices, for example, devices could be setup
Program, Institute of Transportation Engineers
month, day, weekday, hour, minute, and second.
Management and Operations of Intelligent
Transportation Systems (2003). ITS Standards
Overview.
reFerences
IEEE Std 1489-1999 (1999). IEEE Standard for
Aidarous, S., & Plevyak, T. (1994). Telecom- Data Dictionaries for Intelligent Transportation
munications Network Management into the 21st Systems.
Century. New York: IEEE Press.
IEEE Std 1488-1999 (2000). IEEE Standard for
Feit, S. (1995). SNMP: A Guide To Network Man- Message Set Template for Intelligent Transporta-
agement. New York: McGraw Hill, Inc. tion Systems.
ITU-T X.680-X.690 (1994). ISO/IEC 8824: Ab- Steedman, D. (1993). ASN.1: The Tutorial & Ref-
stract Syntax Notation One (ASN.1). erence. London: Technology Appraisals Ltd.
National Highway Institute (2001). Using the
National ITS Architecture for Deployment.
289
290
Chapter 17
Vehicular System
Management Architecture
and Application Platform
Teng-Wen Chang
National Taiwan University of Science and Technology, Taiwan, R.O.C.
Jiann-Liang Chen
absTracT
Notably, not all telematics services can be used in telematics terminals as a result of the varied platform
standards. The main issues are that most telematics technologies depend on vertical, proprietary and
closed per-OEM Original Equipment Manufacture (OEM) platforms, forming islands of non-interop-
erable technology and preventing third-party service providers from creating valuable services. In
this study, the Open Gateway Service Initiative Vehicle Expert Group (OSGi/VEG) was integrated into
an Android platform to generate a vehicular Android/OSGi platform that has the advantages of both
original platforms, such as remote management, rich class sharing, proprietary vehicular applications,
security policies, easy management of application programming interface (APIs), and an environment
with increased openness. Furthermore, this study integrates the cloud computing mechanism into the
Android/OSGi platform, which allows service providers to upload their telematics bundles onto storage
clouds via the provisioning server.
inTroducTion Transport Systems). Telematics is the convergence

of telecommunications and information processing
Over the past few years, with the enormous mar- for automation in cars. So far, quite a number of
ket potentials of telematics industry and the rapid telematics services have been developed by auto-
development information technology, automotive makers and third-party service providers, such as
telematics has been a booming area and indispens- monitoring, emergency roadside assistance, naviga-
able technology, additionally, it also become a hot tion, diver aids, remote diagnostics, entertainment,
R&D area in mobile computing and ITS (Intelligent web browsing, and so on. But not all of telematics
services can be deployed into telematics terminal
DOI: 10.4018/978-1-60566-840-6.ch017 as a result of various standards of platforms. Be-
Vehicular System Management Architecture and Application Platform
sides, the telematics market is immature now. on-line games, surveillance home conditions, and
The main critical issues are most of telematics so on. The multimedia embedded system should
technologies depend on the vertical, proprietary take advantages of high-performance graphic SoC
and closed per-OEM platforms, forming islands or 3D engines when it performed complicated
of no-interoperable technology and preventing operations. The three part of driver aids system
third-party service providers from creating value which ensures security of drivers, economizes the
added services. Consequently, numerous vehicle use of power and so on. The final part of Urban
groups have been some working in establishing Nomadic/Pedestrians Telematics system which
and developing open/standard embedded plat- provides interactive services between users and
forms for vehicles, these platforms contain OSGi/ service providers in non-automotive environment,
VEG, AUTOSAR, AMI-C, CVIS, OSEK/VDX, for example how to find neighbor locations of
Android, and so on. filling stations in urban. The rest of available
Figure 1 illustrates the open Linux operating telematics scenario is that vehicular drivers can
system is ported into embedded on-board termi- download/upload or share information to telemat-
nal, which not only provides a variety of device ics information platforms in Web servers based
drivers such as CAN\LIN\FlexRay car buses, out- on out-networks connection capability.
networks connection modules and so on, and also In this chapter, we will introduce how to com-
offers resources management. The open/standard bine OSGi/VEG into Android platform, making
telematics platforms in telematics middleware new vehicular Android/OSGi platform has both
layer mainly standardize telematics API’s and advantages of original platforms, such as remote
graphic/vocal HMI (human-machine interface), so management/deployment, rich class-sharing,
that both service providers and car manufacturers proprietary vehicular applications, security poli-
can quickly deliver solutions on time to potential cies, and so on.
market and to simplify complexity of development.
Besides, if service providers want to remotely
deploy telematics services to on-board terminal, relaTed works
or road-side centers need to diagnose the situations
of vehicular devices or set-up configurations of google android open platform
telematics applications in terminal, they should
use remote management services rely on open/ Overview of Android Platform
standard telematics platforms.
Telematics applications can be divided into The Android™ delivers a complete set of software
four categories, including VANET embedded for mobile devices: an operating system, middle-
system, vehicular multimedia embedded system, ware and key mobile applications. The Windows
intelligent driver aids embedded system, and Ur- Mobile and Apple’s iPhone now provide a richer,
ban Nomadic/Pedestrians Telematics embedded simplified development environment for mobile
system. The first part of VANET system which applications. However, unlike Android, they’re
makes vehicle can communicate with other ve- built on proprietary operating systems that often
hicles or road-side units via DSRC/IEEE 1609, prioritize native applications over those created
for example if at front of vehicle have accident, by third parties and restrict communication among
it would broadcast emergent message to back applications and native data. Android offers new
vehicles based on VANET embedded system. The possibilities for mobile applications by offering
second part of multimedia system which makes an open development environment built on an
drivers can be able to watch DVB programs, play open source Linux kernel. Real hardware can be
291
Figure 1. Telematics architecuture
accessed through series of standard API librar- in runtime, and native or third-party applications
ies, such as to manage the Wi-Fi, Bluetooth, and in application layer. The Android software stack
GPS devices. As Figure 2 illustrates the Open components are as follows:
Mobile Alliance (OHA) and Google had supported
Android platform, and hope to reach the goal of • Linux kernel: Core services (including de-
ensuring global mobile services to interoperability vice hardware drivers, process and memory
across devices, geographies, service providers, management, security, network, and power
operators, and networks. management) are managed by a Linux 2.6
At this stage, Google had released the open kernel. The kernel also provides an abstrac-
source of Android platform, providing the option layer between the hardware and the rest
portunity to create new adaptive mobile platform of the stack.
interfaces and applications designed to look, feel • Libraries Running on top of the kernel,
and function exactly as our imagination. Conse- Android includes various C/C++ core li-
quently, the Android platform had been ported braries such as libc and SSL, as well as:
in mobile devices, such as notebook, PDA, and ◦ A media library for playing multime-
automotive system in recently. In automotive dia resources.
system fields, the Intel and Wind River Sys- ◦ A Surface manager to provide display
tems companies are actively working on getting management
Android-powered infotainment operating system ◦ Graphics libraries that include SGL
integrated into vehicles. and OpenGL for 2D and 3D graphics
◦ SQLite for data storage
Android Software Stack ◦ SSL and WebKit for integrated web
browser and Internet security
The Figure 3 illustrates the Android software • Android runtime: Including the core li-
stack, which composed of Linux kernel, a collec- braries and the Dalvik virtual machine, the
tion of Android libraries, application framework Android runtime is the runtime environment
that provides management of android application for android application to be run in normal.
292
Figure 2. Open environment for android platform
• Core libraries: While Android develop- available in the core Java libraries as well
ment is done in Java/C++, Dalvik is not a as the Android-specific libraries.
standardized Java VM. The core Android • Dalvik virtual machine: Dalvik is a
libraries provide most of the functionality register-based virtual machine that’s been
Figure 3. Architecture of android software stack
293
optimized to ensure that devices run multi- Features of Android Platform

ple instances efficiently and stable. It relies
on the Linux kernel for threading and low- The unique features which other mobile platforms
level memory/process management. don’t own as follows:
• Application framework: The applica-
tion framework provides the classes used • Google map of navigation
to create Android applications. It also pro- applications:Figure 4 illustrates Android
vides a generic abstraction for hardware platform provides the MapView widget lets
access and manages the user interface and navigation display, manipulate, and anno-
application resources. tate a Google Map within your Activities
• Application layer: All applications, both to build map-based applications using the
native and third party, are built on the ap- familiar Google Maps interface.
plication layer using the same API librar- • Background service and applications:
ies. The application layer runs within the Background services let android appli-
Android runtime using the classes and ser- cations that use an event-driven model,
vices made available from the application working silently while other applications
framework. are being used and in foreground.
• All applications are created equal:
Android does not differentiate between the
core applications and third-party applica-
Figure 4. Location-based services for android tions. They can all be built to have equal
platform access to a platform’s capabilities provid-
ing users with a broad spectrum of applica-
tions and services.
• P2P interdevice application messaging:
Android offers peer-to-peer messaging
mechanism that supports presence, instant
messaging, and inter-device/inter-applica-
tion communication.
• Breaking down application boundaries:
Android breaks down the barriers to build-
ing new and innovative applications. The
user can combine information from web to
his/her individual’s Android platform, such
as the user’s contacts, calendar, or geo-
graphic location to provide a more relevant
user experience. Besides, all of the third-
party applications can be downloaded to
user’s Android platform from Android
Market on free.
294
open services gateway initiative loadable java-based service applications known as

(osgi) service platform bundles. OSGi bundles are applications packaged
in a standard Java Archive (JAR) file, containing
Overview of OSGi Framework a manifest file that describes the relationships
among bundles and OSGi framework, a series of
The OSGiTM specifications define a standardized, compiled Java classes, and native code. Figure
components-oriented, computing environment for 5 illustrate the OSGi framework run on top of a
networked services that is the foundation of an Java virtual machine, and the framework provides
enhanced service-oriented architecture (SOA). a shared execution environment that installs, up-
Besides, the OSGi specifications were initially dates, and uninstall bundles dynamically without
targeted at residential internet gateway with home restarting the system.
automation applications (Lee et al., 2003; Myoung
et al., 2005); however, the OSGi features of the OSGi-Based Telematics Gateway
standard and extensible that making enterprise
to regard OSGi technology as key solutions. For The OSGi Alliance whitepaper had pointed out
example, Nokia and Motorola drove the OSGi that the research and development of OSGi have
technology standard for the next generation of extended to any networked environment, such
smart phone, and the vehicle industry adopted the as automobile, which provided some complex
OSGi specifications by embedding them into the embedded devices and the requirement of de-
global system for telematics (GST) specifications, ployable services for those devices. The OSGi
which is supported by many car manufactures. framework make it possible to access, download
The OSGi framework provides general-purpose, and use services in automotive mobile environ-
secure support for deploying extensible and down- ment, creating an enormous potential for automo-
Figure 5. Architecture of OSGi framework
295
tive manufacturers and service providers to offer • Driver aid bundle: It can provide driv-
new proactive services to drivers and passengers er assistance for vehicles, and offering
(Gun et al., 2001). The open standard-based, some warning message to driver by lane
service-oriented automotive infrastructure for the recognition.
telematics architecture(Li et al., 2005; Zhang et • Diagnostics bundle: It can provide remote
al., 2004), including OSGi gateway, service infra- diagnostic services, and complicated diag-
structure and self-automobile. The central control nostic information also can be analyzed and
point of the service infrastructure, which is the stored in remote systems, in this technique
OSGi-based telematics gateway, can be intercon- is similar with MYCAREVENT project
nected with various internal and external networks (Weiss et al., 2006), making the telematics
and buses. It acts as the execution environment gateway to have the capacities of automat-
for mobile automotive services. The following ic repair and self-management.
bundles provide great deal of critical services for • Information/entertainment bundle:
OSGi-based telematics gateway: At present, video and audio equipments
have been widely applied in vehicles.
• Monitoring bundle: It is the one of basic Consequently, existence of researches that
intelligent vehicle functions. Though mon- will shift to provide an infotainment server
itoring state and performance of local vehi- system for the in-vehicle users such that
cle devices, driver can acquire information the network-enabled information appli-
about vehicles, such as temperature, and ances (IA) can access the information and
pressure of types, engine levels, etc. perform the entertainment services from
• Navigation bundle: It can provide navi- infotainment server (Hsu et al., 2005).
gation system that offers navigational as-
sistance to drivers. The system receives The automotive operating system is also more
GPS position information signals which significant to the vehicles (Ai et al., 2007), by
are processed to determine current position supporting specific device drivers such as CAN/
latitude and longitude coordinates, more- LIN buses, which is used to communicate to other
over the direction of travel. electronic control units (ECU) nodes for diagnostic
• Communication bundle: It can provide purpose, consequently existence of researches that
OSGi-based telematics gateway to connect embedded OSGi platform to the vASOS automo-
with in-vehicle or out-vehicle network. tive operating system (Sun et al., 2007). Figure 6
Besides, this bundle can make OSGi-based illustrates the OSGi-based middleware run on top
gateway to be extended its functionality of the K virtual machine (KVM), which is a com-
by remote deployment of OSGi bundles pact, portable Java virtual machine intended for
that service provider provided. Through small, resource-constrained devices. In addition,
CAN, LIN or MOST(Zhou et al., 2006), those researches choose OSGi R3 implementation
the OSGi-based telematics gateway realiz- of Oscar, to develop its application bundles, such
es local area connection among in-vehicle as navigation and CAN/LIN/MOST access bundle.
devices, and by using GPRS, GSM or even In our studies, we had took advantage of OSGi
special ITS FM channels, it realizes wide R4 implementation of Apache Felix to integrate
area connection among vehicles, home with Google Android platform, making a variety
gateway, remote service center, and road- of Google APIs to turn into OSGi bundles, such as
side systems. location-based, peer-to-peer communications, and
296
Figure 6. Architecture of vASOS and OSGi-based middleware for vehicles
network manager, etc. Not only supplying service service delivery. More specifically, 3GT
providers to deploy its telematics services, but also will develop OSGi-based specifications for
enhancing Android runtime layer performance and the interface between vehicles and control
remote management functionality. centers as well as for the interface between
The goal of the Vehicle Expert Group (VEG) control centers and service providers.
is to tailor and extend the OSGi specification in • ITEA EAST EEA: This automotive proj-
order to meet vehicle-specific requirements. To ect is funded by the European Union and
achieve this, the VEG will define a list of topics consists of major European automotive
that cover vehicle-specific issues. Together with manufacturers, first-tier suppliers and re-
the automotive industries, the VEG will specify search departments. The goal of EAST-
corresponding telematics API’s, so that both ser- EEA is to enable hardware and software
vice provider and car manufacturers can quickly interoperability of in-vehicle ECU nodes
deliver solutions on time to market and to increase through definition of an open, middleware-
customer loyalty. The OSGi-VEG organization based architecture.
has the following automotive projects: • Stadtinfokoeln: Stadtinfokoeln is a
Cologne-based project, funded by the
• 3GT: The 3GT project is to help establish German Ministry of Education and Science
OSGi-based in-vehicle telematics plat- (BMBF), and focused on the delivery of
forms on the European mass market by en- parking services to the inhabitants or visi-
suring interoperability between the prod- tors of Cologne, Germany. The goal is to sat-
ucts of different middleware providers, isfy customer demand and to reduce traffic,
terminal manufacturers and service provid- which is related to the availability of park-
ers. This will be done by establishing coming space. An important role for this is the
mon telematics interfaces for OSGi-based rapid development of the information and
297
Communication-Technology in the mobile high-cost of consumption, they maybe encounter

environment. To fulfill all requirements the the hazardous conditions. Therefore, in open en-
decision was made to use an OSGi service vironment it is hard to establish meaningful trust
platform framework on the vehicle plat- relationships, and even when one can, trust is not
form. This enables the rapid development equated with quality. The existence of research
and deployment of new services. which utilizes the Aspect-Oriented Programming
• TOP-IQ: The Eureka-project TOP-IQ (AOP) to implement security enforcement policy
(Hackbarth, 2003) focuses on the develop- to weave into OSGi bundles (Phung et al., 2008).
ment of a new generation of on-board com- Additionally, the AOP provide the cross-cutting
puters (OBC) for luxury cars and trucks. functionalities into complicated vehicular ap-
The OBC facilitates telematics services for plications, and program concerns in a software
transport, logistics, offers the management, system can be captured and encapsulated in to
billing, and delivery of new services such so-called aspects. Figure 7 illustrates the weaving
as online navigation, trailer management/ process scenario in vehicular environment, which
tracking and tracing of cargo, etc. consists of in-vehicle system, control center, and
third-party service providers.
Safe Deployment of OSGi Bundles The vehicle owners would like to install vehicu-
lar application bundle by making a corresponding
The Security of deployment vehicular services is request in the GST standard. Before installing
crucial issues in telematics environment, suppos- this application bundle on the in-vehicle system
ing vehicle owners would like to install third-party through wired/wireless network, the control center
applications to their on-board vehicle system, but weaves the application with defined aspect such
if those vehicular applications have many bugs, as security policies, so that the woven bundle can
such as sending mass messages, accessing the be operated with enforcement security policies
security sensitive parts of vehicles, and causing into the in-vehicle system.
Figure 7. Weaving process in vehicular environment
298
comparions oF osgi Consequently, OSGi services play significant role

and android in keeping components lightweight and loosely
coupled. In OSGi specification, it defines many
Figure 8 illustrates both OSGi and Android define of remote management services, such as SNMP,
a development platform for developing service- CMISE, CIM and OMA DM (Hyun et al., 2008).
oriented computing mechanism in mobile devices, By utilizing remote management services, we can
and existence of similar VM technology, service easily remote-manage OSGi framework in many
model, component model and feature of open. But situations, but these services are non-available
the main difference is that Android platform uti- in Android platform. Android platform lack of
lizes process-based isolation to enhance resource class-sharing ability seriously limits the sharing
management, however OSGi platform takes capa- of components and remote device management.
bility of classloader-based isolation model, which The lack of versioning and lack of management
allows OSGi bundles to selectively export some of APIs is a further limitation. Presumably, OEMs
their internal packages to other required bundles. will provide some solutions for these important
Bundles can declare their package dependencies requirements.
and the framework is responsible for managing
dependency relationship between required bundles
and provided bundles. Based on classloader- open embedded android/osgi
based isolation mechanism, OSGi framework plaTForm For TelemaTics
has advantage of rich class sharing. On the other
point of view, Android platform uses background location of osgi Framework
services to perform time-consuming operations, in android platform
but usually these services are heavyweight, and
only can be accessed by utilizing inter-process The Dalvik VM use the device’s underlying Linux
communication, which makes service access kernel to handle low-level functionality including
slower and more expensive. In OSGi, service security, threading, process, and memory man-
components are lightweight. They are called agement. It’s also possible to read/write C/C++
directly when finding them in service registry. applications that run directly on the underlying
Figure 8. Comparisons of OSGi and android
299
Figure 9. Location of OSGi framework

Linux OS. But Dalvik VM merely can execute
Dalvik executable files (e.g., classes.dex), a
format optimized to ensure minimal memory
footprint; consequently it can’t perform general
Java packages as well as OSGi bundles, so that
OSGi framework operations to fail when Android
realizes OSGi functionalities in runtime. Besides,
as Figure 9 shows, if the OSGi framework success
to run in Android Runtime layer, it can’t directly
interact with users, because users only can catch the
sight of Android Activities (same as application)
in application layer. For this reason, the Android/
OSGi Activity must to grab the report or log mes-
sage of OSGi framework, and should to implement
Android UI to present information to the users. We
consider the OSGi R4 implementation of Apache
Felix to integrate with Android platform, because
this container is more portable and light-weight,
and being appropriate for embedded hardware
device such as telematics system.
android/osgi conversion Figure 10. Android/OSGi conversion mecha-

mechanism nism
The Android/OSGi conversion mechanism can

assume OSGi framework to dynamically load-
ing bundles through Android Davik VM. In
first, as the Figure 10 shows the Android/OSGi
conversion mechanism, which can make OSGi
framework perform a series of complied Java
classes from OSGi bundles in Android runtime
layer, by means of classes.dex file to referencing
OSGi bundle’s general Java classes. Therefore,
the classes.dex file can be regarded as bridging
role between OSGi framework and bundles. We
had implemented Android/OSGi conversion code
through Dalvik.system.dexFile API, and import-
ing two parameters, such as package name(bundle
API packages) and ClassLoader (reading the Java
classes) objects, and then conversion code will
return the collection of Java classes, so that OSGi
framework can obtain this collection objects to
indirectly perform the functionalities that OSGi
bundles provided.
300
After accomplishing the Android conversion by means of Dalvik VM reading the main class
mechanism, we will make sure all OSGi bundles of OSGi framework, afterward will enter OSGi
or general Java packages to contain classes.dex console mode, and showing list of some installed
files, so that OSGi framework can assume any bundles. Nevertheless, in this technique, the users
bundle to work, making API packages to share, can’t directly interact with OSGi framework in
and letting consumer bundles to query the OSGi real hardware device, so we had implemented An-
services that producer bundles provided. We had droid/OSGi Activity and GUI interface, to acquire
adopted the DEX conversion tool that Android corresponding OSGi information from underlying
SDK provided. This tool can easily make Java layer, and also embedding OSGi framework into
complied file to have classes.dex file that Dalvik Android application that can enhance it to be more
VM is able to execute. Figure 11 illustrates the adaptive and powerful.
conversion flow of OSGi bundles to become apk
file (Android executive extension); in originall, we development of android/
had made the OSGi bundle projects to be compiled, osgi activity
and if no exception will generate a series of Java
compiled classes. After making complied classes We had created Android/OSGi Activity in Android
to come into being jar file, we took DEX tool to application layer, which can fetch the OSGi corre-
make jar file containing classes.dex, and then by sponding information from Android runtime layer,
using aapt command to transform OSGi bundles and providing interactive environment between
into Android apk files. In the end, by using adb users and OSGi framework. In first integration of
push command to push converted apk files to procedure, we must to implement Android Acitiv-
Android platform. ity interface, letting our Android/OSGi application
With Finishing Android conversion mechanism to have Android Activity properties. Unlike most
and converted OSGi bundles, we had ported OSGi traditional environment, Android applications
framework to Android underlying layer, and other have no ability of control over their own life
functional bundles such as telnet, deployment cycles, Instead, Android application framework
admin, http, and remote manager bundles etc, so must to be in charge of android applications state,
that Android platform can be assisted and managed and react accordingly, taking particular care to
in remote site that original platform doesn’t have be prepared for untimely termination. When
this ability. Figure 12 shows the script of start- Android/OSGi Activity becomes active state and
ing OSGi framework in Android runtime layer, need to operate continually, the Android applica-
Figure 11. Conversion flow of OSGi bundles to become apk files
301
Figure 12. OSGi framework run in android runtime layer
tion framework will pay attention to ensure that the main class of OSGi framework after Android/
the Android/OSGi Activity remains responsive. OSGi Activity onCreate() method had already
Therefore Android application framework has been initialized. Android services run without
the capacity of monitoring platform resources at a dedicated GUI, they usually silently perform
any time, if necessary, will kill some empty or background works that users can’t perceive. Before
background processes to free resources for high- Android services to run, they must to be attached
priority applications. with our Android/OSGi Activity. From different
The Android application is called Activity, point of view, the Android/OSGi Activity provides
which represents a screen that an application Android Context object, which makes Android
can present to its users. Android Activity typi- application framework to be able to manage this
cally includes at least a primary interface screen Activity life states, and communication between
that handles the main UI functionality for our Android/OSGi Activity with other native or third-
application. This is often supporting additional party Activities.
by secondary Activities for entering information, To ensure that our Android/OSGi Activity
and providing different perspectives on our data. remains responsive, we move all slow, and time-
For example, we had created the Android views consuming operation off the main Activity thread
of OSGi console, making users can enter OSGi and onto a child thread, that is continually reading
commands and catch the sight of OSGi corre- the main class of OSGi framework, and updating
sponding Information. Every Android Activity GUI interface when OSGi console view has been
has an initializer method called onCreate(), by changed or GUI bundles provide the new screen
using this initializer method, we can distribute the services. Consequently, we had implemented the
memory to perform the launcher thread which can Felix service to do the above-mentioned opera-
make OSGi framework to be started. Figure 13 tions, as Figure 14 shows that Felix service can
illustrates the cross-communication operation that directly find out the location of OSGi framework
Android/OSGi Activity launch the FelixService in Android runtime layer, and extract the main jar
(Android background service) which can represent file of OSGi framework, which is Apache Felix
the GUI interface of OSGi console and to load of R4 OSGi implementation, and by means of
302
Figure 13. Cross-communication operation
Dalvik ClassLoader object to read the main class embedding osgi Framework
continually in child thread, so that we can start OSGi to android application
framework in Android application layer. In addition,
we also had implemented the OSGi console view In traditional, the Android/OSGi platform utilizes
that can fetch the OSGi corresponding message, the OSGi services and service registry as the ex-
and represent to users. The OSGi console view tensibility mechanism. This mechanism will let
has two input/output interfaces that make users to any application to run completely on top of OSGi
immediately interact with OSGi framework. framework, but this is not always possible, in other
Figure 15 illustrates two-way Communication words, the components of system should follow
Mechanism between OSGi and Android. The the OSGi standard to be implemented, which
Android/OSGi Activity(described in Figure 13) will result in increasing the development times
can transmit these received information to the and costs. Besides, it should spend extra-time to
OSGi framework after finishing above-mentioned communicate with OSGi framework in runtime
operation(First step of Figure 15, sending data such layer. Therefore, if any Android application wants
AP table), so that OSGi framework can regard to use either the OSGi services or provided APIs
these information as permanent properties data by bundles, it needs complicated Android/OSGi
(key/value pairs) stored in OSGi environment(Step communication mechanism. We had created ex-
2 in Figure 15), and OSGi bundles can retrieve tender mechanism that makes OSGi framework
these properties through preference service to to tightly embed into Android applications in
implement advanced applications such as network application layer as Figure 16 shows, in this way,
manager, and navigation bundle, etc(Step 3 in any android application can host the instance of
Figure 15). In the same way, when OSGi bundles OSGi framework by utilizing reflection approach,
becoming ACITVE state, it also can transmit and application can externally load the services
some explicitly or implicitly intents to launch which OSGi bundles provided. Nevertheless, the
useful Activities. Consequently, on the strength extender mechanism has some drawbacks about
of this two-way communication mechanism, we lack of dynamic changes in OSGi bundles/services
can make Android applications to connect with and configuration of OSGi instance. Therefore,
OSGi framework and bundles easily. we had combined two potential mechanisms with
303
Figure 14. Inter-architecture of the Felixservice
Figure 15. Two-way communication mechanism between OSGi and Android
Android platform via cross-communication inter- interface to monitor services. As a result of utiliz-
face between felixService and OSGi Framework, ing traditional service-oriented mechanism, the
management between framework layer and em- producer bundles should use its BundleContext
bedded OSGi Activity, to provide more integrated object to register services in service registry,
Android/OSGi environment. and if services referenced by consumer bundles
The Felix framework instance doesn’t utilize had been changed, we must implemented Ser-
ServiceReference array object to get referenced viceChanged event to be careful of no exceptions
services, but on the contrary it use Service Tracker had been generated, hence this mechanism is more
304
Figure 16. Tightly integrated Android/OSGi environment
inconvenient than service tracker mechanism. The Vehicular android/

service tracker takes advantages of whiteboard osgi applicaTions
pattern, making embedded OSGi framework
to utilize server bundle to monitor referenced Vehicular Testbed
services changes in central way. For example,
as Figure 17 shows, if referenced services had It is inconvenient and high-cost to test our Android/
been modified, the server bundle automatically OSGi applications to real vehicles, consequently
callback modifiedService method to rebind these as Figure 18 illustrates we build a vehicular testbed
services without implementing ServiceChanged to test and verify the functionalities of Android/
event handler. For example, the embedded OSGi OSGi applications. Besides, we regard sensor-
Activity has capacity of tracking UI services, based LEGO robots as simulated vehicles, and
and if UI services have existed, the embedded vehicular Android/OSGi platform acts as on-board
OSGi Activity would utilize child thread to call terminal for simulated vehicles. Instead of utilizing
setContentView method to update the foreground car buses, this on-board terminal mainly controls
screen to users. simulated vehicles through Bluetooth like as IBM
Figure 17.Service tracker of embedded OSGi activity
305
iDrive systems. And communications can be done in order to lighten the use of power. The second
via Wi-Fi among different vehicles. part of application is making vehicles to have
visual intelligence that can identify objects like
android/osgi applications as traffic signals. The rest part of application
is making vehicles to keep distance with front/
We developed various telematics services to back vehicles. Android/OSGi applications can
establish intelligent vehicles in mobile environ- be remote-deployed into on-board terminal by
ment, which mainly consist of line follow, object automaker or service providers, but if some of
detection, keep distance, and so on. As Figure applications access the significant components
19 shows the first part of application is making of vehicle, the vehicular Android/OSGi platform
vehicles to follow guideline to ensure security can enforce security policies to avoid accidents
of drivers and no necessary to finding your path through AOP-based OSGi weaving mechanism.
Figure 18. Vehicular testbed
Figure 19. Android/OSGi applications
306
Original Android platform can make developer remote management

to create map-based Activities using Google Map
as a user interface element. We have full permis- Figure 21 illustrates vehicular Android/OSGi
sions to access the map, include change the zoom platform mainly provide three management con-
level, move the centered location, use overlays, soles, which include telnet, web management, and
provide map-contextualized information and remote deployment console. The first part of telnet
functionality, and so on. But if service providers console can make remote manager to diagnose/
want to design navigation applications based on manage Android/OSGi applications though telnet
map-based Activities, they must to create many interface, which extends telnet bundle that Oscar
heavyweight services to perform time-consuming R3 implementation provided. Original Android
operations. Besides, Android platform lack of platform can’t utilize telnet services to diagnose/
class-sharing among different applications and manage its conditions of applications or system
islands of no-interoperable processes, which of- configuration. The second part of web console
ten limit usage of system resources and remain management can make remote manager to diag-
non-responsive. Figure 20 illustrates vehicular nose/manage Android/OSGi applications based
Android/OSGi platform has feature of lightweight on web techniques. The web management console
applications, which application size is less 4KB has been established rely on underlying OSGi
than original application. But vehicular Android/ bundles, which contain http service, declaractive
OSGi platform must to spend about 3 seconds to service, log service, configadmin, deployadmin,
start OSGi framework in first time, and additional metadata, and so on. Although original Android
memory size of OSGi framework. platform has http service, that can give users
Figure 20. Lightweight Android/OSGi applications
307
Figure 21. Remote management of vehicular Android/OSGi platform
way to surf internet depend on WebKit browser, reFerences

it doesn’t have web server, which allow remote
users to login/logout into platform. The rest part Ai, Y., Sun, Y., Huang, W., & Qiao, X. (2007).
of remote deployment console can make applica- OSGi based integrated service platform for
tions be able to be deployed into remote vehicular automotive telematics. In Proceedings of IEEE
Android/OSGi platform. This way is similar with International Conference on Vehicular Electronic
Android market deployed mechanism, but has and Safety, (pp. 1-6).
something differences, such as original Android Gun, N., Held, A., & Kaiser, J. (2001). Proac-
platform can’t administrate the identity of users tive services in a distributed traffic telematics
and automatically validate security and stability application. International Workshop on Mobile
of applications. Communication over Wireless LAN: Research
and Applications.
308
Hackbarth, K. (2003). OSGi – Service-Delivery- Phung, P. H., & Sands, D. (2008). Security Policy
Platform for Car Telematics and Infotainment Enforcement in the OSGi Framework Using
Systems. Advanced Microsystems for Automo- Aspect-Oriented Programming. In Proceedings of
tive Applications 2003 (pp. 497-507). Berlin: the IEEE International Conference on Computer
Springer. Software and Applications, (pp.1076-1082).
Hsu, R. C., & Chen, L. R. (2005). Integrated Sun, Y., Huang, W. L., Tang, S. M., Qiao, X., &
Embedded System Architecture for In-Vehicle Wang, F. Y. (2007). Design of an OSEK/VDX
Telematics and Infotainment System. Proceedings and OSGi-based embedded software platform for
of the IEEE International Symposium on Industrial vehicular applications. In Proceedings of IEEE
Electronics, 4, 1409–1414. International Conference on Vehicular Electronic
and Safety, (pp.1-6).
Lee, C., Nordstedt, D., & Helal, S. (2003). En-
abling Smart Spaces with OSGi. IEEE Pervasive Weiss, E., Gehlen, G., Lukas, S., & Rokitansky,
Computing, 2(3, July-Sept), 89-94. C. (2006). MYCAREVENT- Vehicular Com-
munication Gateway for Car Maintenance and
Li, Y., Wang, F., Feng, H., & Li, Z. (2005). OSGi-
Remote Diagnosis. In Proceedings of the IEEE
based service gateway architecture for intelligent
International Conference on Computers and
automobiles. In Proceedings of the IEEE Interna-
Comunication, (pp. 318-323).
tional Conference on Vehicles Symposium, (pp.
861-865). Zhang, D., Xiao, H. W., & Hackbarth, K. (2004).
OSGi based service infrastructure for context
Myoung, K., Heo, J., Kwon, W. H., & Kim, D.
aware automotive telematics. In . Proceedings
S. (2005, July). Design and implementation of
of IEEE International Conference on Vehicular
home network control protocol on OSGi for home
Technology, 5, 2957–2961.
automation. In . Proceedings of the IEEE Interna-
tional Conference on Advanced Communication Zhou, Y., Wang, X., & Zhou, M. (2006). The
Technology, 2, 1163–1168. Research and Realization for Passenger Car CAN
Bus. In Proceedings of the IEEE International
Forum on Strategic Technology, (pp. 244-247).
309
310
Chapter 18
Remote Vehicular System
Management Functions and
Information Structure
Teng-Wen Chang
Jiann-Liang Chen
absTracT
Due to the rapid development of information technology, the network has already spread to every corner of
vehicle. With all kinds of ECU devices appear in the vehicle, and it brings the more and more convenient
living. On purpose solving heterogamous technologies that are incompatible with each other, developed
a “WBEM-based Remote Management and Heterogeneous Vehicular Network Diagnosis System” on
OSGi Gateway. This system can focus on a variety of problems come from vehicle network, and find
out what are the problems or where are the problems happened. If the problem still can not be solved
properly, we must to seek for help from remote managers. The users can acquire enough information
without understanding how to control every device, so that the users can help near diagnosis system to
solve vehicle network’s problems and to promote the abilities of near network diagnosis.
inTroducTion constructs a remote vehicular network management

and diagnosis system with OSGi platform (Ai et al.,
Due to the vehicular network management system 2007; Sun et al., 2007) and WBEM (Web-Based
has not been standardized and lacking interoperabil- Enterprise Management) to speed up the realiza-
ity. In addition to the existing network management tion of the heterogeneous home network. Usage
system has a simple structure consisting of server, of OSGi platform provides several benefits. The
provider and client, as communication between most important one that it integrate different types
admin users and management server. This study of vehicular networking technologies, because cur-
rently most available vehicular network middleware
and technology have no compatibility with each
DOI: 10.4018/978-1-60566-840-6.ch018
Remote Vehicular System Management Functions and Information Structure
Figure 1. Remote managing protocols
other, home devices based on these heterogeneous ment. Using WBEM protocols, remote managers
middleware cannot communicate with one another, can easily manage devices through web interfaces,
even though they are physically connected. and operate with any manufacturer’s device, re-
The OSGi service platform is initially chosen gardless of the underlying protocols.
for its capabilities to integrate components and The proposed system uses the WBEM to pro-
services from different providers in the hetero- vide more effective resource management and a
geneous home network. (Marples et al., 2001; larger range of services than before. Because the
Saito et al., 2000) The OSGi service platform is different types of vehicular network technologies
specifically designed for devices that can manage and management instruction, need the different
remote devices through remote managers. Figure technology to manage the each vehicular network,
1 illustrates these devices that need to be estab- like SNMP, Telecommunications Management
lished some of remote managing protocols, such Network (TMN) and so on. In order to overcome
as SNMP, CMISE, CIM, OMA DM, and more. the situation that WBEM and CIM (common
The OSGi Alliance decided that no managing information model) provide an excellent manage-
protocol can be preferred over others because ment environment and promotes the information
no protocol is suitable for all cases. The OSGi exchange across a variety of underlying technolo-
Alliance now has been working to develop a stan- gies and platforms supported interoperability. The
dardized management API to build functionality WBEM offer extreme flexibility and efficiency
of remote management into unattended devices. to manage method, that communication with
This is a very powerful concept that offers the each other different management type. Figure 2
same interoperability as a standard protocol. How- illustrates overview of integrating OSGi platform
ever, the benefits of this concept are not always with WBEM these two technologies and Remote
immediately obvious. The benefit of a standard management through Web-based Enterprise Man-
protocol is that any device can be managed by agement Architecture to the vehicular gateway.
any manage operators. Consequently, we apply The WBEM-based Remote Vehicular Network
Web-based Enterprise Management (WBEM) Management and Diagnosis System can figure out
managing protocol to vehicular network environ- the status fast and easily, and realize the benefits
311
Figure 2. Integrating OSGi platform with WBEM
of integrating these two technologies, significant applied widespread in industrial automation and
research has been carried out the telematics in automotive/truck applications. The FlexRay pro-
the near future. tocol provides high-speed, deterministic and fault
tolerant communications for in-network. It also
can be compatible with existing networks, such
relaTed works as CAN, LIN, and so on. But above-mentioned
protocols can’t intercommunicate to each other.
We make a brief introduction of vehicular network Consequently, the OSEK/VDX operating system
system which are popular nowadays, and make acts a role which makes the message communi-
use of scenario situation to state vehicular network cate between two different car protocols (Kim et
environment in the future, state a lot of technol- al., 2007).
ogy applied to vehicular network in this chapter. In the out-vehicle network, the OBU (On
Besides, related work would also be discussed. Board Unit) in the vehicle can communicate to
infrastructure via out-network modules, or con-
Vehicular network environment nect with another vehicles via DSRC/IEEE 1609.
The remote home service provider and remote
Figure 3 illustrates the vehicular network environ- vehicular service provider can provide its par-
ment, which can be divided into two categories, ticular services to automotive user.
including in-vehicle network architecture and
out-vehicle network architecture. In the in-vehicle open services gateway initiative
network, the vehicle might include four kinds of (osgi) service platform
car bus protocols such as CAN (Controller Area
Network), LIN (Local Interconnect Network), The OSGi™ Alliance was founded in March 1999.
MOST (Media Oriented Systems Transport) Its mission is to create open specifications for the
and FlexRay (Mariño et al., 2009). The CAN network delivery of managed services to local
bus protocol is main backbone for in-network networks and devices. The OSGi organization is
environment in current, and it has the goal of the leading standard for next-generation Internet
making vehicle more reliable, safe and fuel- services to homes, cars, mobile phones, desktops,
efficient while decreasing wiring harness weight small offices, and other environments. With the
and complexity. Besides, CAN bus protocol has advent of ubiquitous communication, PCs as well
312
Figure 3. Vehicular network environment
as diverse devices, such as sensors and information manages the dependencies between bundles and
home appliances, are now being linked together, services in detail. It provides the bundle developer
often using heterogeneous communication proto- with the resources necessary to take advantage
cols. In addition, broadband access has become of Java’s platform independence and dynamic
much more widespread and it is common to see code-loading capability in order to easily develop
homes and offices with always-on connections to services for small-memory devices that can be
the Internet. Against this backdrop, increased at- deployed on a large scale. The OSGi framework
tention is being paid to gateways that provide key architecture has a number of layers as depicted
capabilities for functional interworking between as Figure 4.
devices and portal capability for using services of-
fered by external networks, including the Internet. web-based enterprise
The OSGi Service Platform is delivered in many management (wbem)
Fortune Global 100 company products and ser-
vices and in diverse markets including enterprise, WBEM stands for Web-Based Enterprise Manage-
mobile, home, telematics and consumer. ment. WBEM is an ongoing initiative started by in-
The core component of the OSGi specifications dustry leaders such as Compaq, Microsoft, Cisco,
is a framework. It provides a general-purpose, se- Intel and over 70 others. This initiative proposes a
cure, and managed Java framework that supports common method of managing enterprise systems.
the deployment of extensible and downloadable It merges existing management solutions with the
applications known as bundles. OSGi-compliant latest advances in Web technology. WBEM is a
devices can download and install OSGi bundles, set of Internet standards which gives the ability
and remove them when they are no longer required. to interconnect between different management
The Framework manages the installation and standards and environments. WBEM allows
update of bundles in an OSGi environment in a managing both software (operating systems, ap-
dynamic and scalable fashion. To achieve this, it plications) and hardware (computers, network
313
Figure 4. OSGi framework architecture
devices) by creating a common layer which uni- The CIM specification is the language and
fies and simplifies management through WBEM methodology for describing management data.
compliant applications. CIM is an object-oriented schema for modeling
WBEM is an initiative of DMTF and it includes the objects. The CIM schema can be divided to
a set of technologies that enables the interoper- three areas; the core model, the common model
able management of an enterprise network. The and the extension model. First, the core model
DMTF has developed a core set of standards that captures notions that are applicable to all areas of
make up WBEM, which includes a data model, management. Second, the common model is an
the CIM standard; an encoding specification, information model that captures notions that are
CIM-XML encoding specification; and a transport common to a particular technology. For example,
mechanism, CIM operations over HTTP. Figure it includes the model for systems, applications,
5 illustrates the relationships among WBEM networks and devices. Last, the extension model
Standard Technologies architecture. represents technology-specific extensions of
common models.
Figure 5. Relationships among WBEM standard technologies
314
Figure 6. WBEM Architecture

The CIM-XML encoding specification defines
XML elements, written in Document Type Defi-
nition (DTD) that can be used to represent CIM
classes and instances. The CIM operations over
HTTP specification defines a mapping of CIM
operations onto HTTP that allows implementations
of CIM to interoperate in an open, standardized
manner and completes the technologies that sup-
port WBEM. Therefore, in the WBEM architec-
ture, the management information is described by
the CIM schema, converted to XML, and finally
embedded in an HTTP payload to transport to
the target node.
Figure 6 illustrates the WBEM architecture
(Thompson et al., 1998), which includes a WBEM
Client, and WBEM Server. WBEM Server has
CIM Object Manager (CIMOM) which is a central
component that routes information about objects
and events between components. The CIMOM
responds to the operations defined in the “CIM
operations” specification such as create, modify,
and delete. It also checks the syntax and semantic
be hard-coded, table-driven or driven from the
of the messages, and provides security. Providers
model.
are so-called instrumentation agents. Namely,
Once the actual CIM commands have been de-
Providers actually obtain the information from the
termined, they are encapsulated in CIM-XML and
resources and forward it to the CIMOM. A WBEM
passed to the HTTP client. This is responsible for
Client is commonly represented as the manage-
any negotiation required with the WBEM server
ment application, and it can get the information by
and the correct transfer of the request. Addition
sending a request message to the CIMOM instead
to traditional CIM operations using XML, CIM
of directly accessing the providers.
operations using Binary XML (Park et al., 2006)
Figure 7 shows the various layers through
enhance a performance of message transportation
which an operator’s request passes and through
between WBEM clients and a WEBM server.
which the response returns when a managed object
Once the request reaches the WBEM server,
is accessed.
it is passed from the HTTP server to have the
A command from the operator or higher-level
CIM-XML reconstituted into CIM commands. The
management system originates with the applica-
CIM Object Manager (CIMOM) examines these
tion logic which handles graphical screens, in-
and determines how they should be handled: by
terpretation of command-line interfaces, etc. The
the CIMOM itself, with reference to the schema
command is then mapped to the abstract object
stored in the repository, or by a provider. If it is
model. Fundamentally, the Object Abstraction
passed to the provider, then the provider accesses
layer has to know that, when the operator enters
the real device or service to retrieve the requested
a command to create a new OSPF service, for
information or carry out the requested configura-
example, then this means the creation of instances
of various CIM classes. This knowledge may
315
Figure 7. WBEM Client/WBEM server interactions
tion. The response from the provider follows the • Standard Interface for different moni-
reverse path back to the operator. toring systems: Much fabric management
WBEM is architecture, not an implementation. uses the WBEM standards.
Implementations exist though. Sun Microsystems • Standard Interface for information
has released a Java WBEM and implemented providers: Future APIs available for
CIMOM with Java Management Extensions and simplicity.
Microsoft has its CIMOM implementation called • Can replace log file style providers:
Windows Management Instrumentation (WMI). Provide CIM repository.
Another similar SDK that facilitates development
of WBEM servers is the SNIA CIMOM, a free Replacing existing management standards such
product with open source code that is compatible as SNMP or CIMP is not WBEM attempt to, but to
with the Sun WBEM SDK API. Table 1 shows provide a framework embracing existing manage-
existing WBEM implement. ment standards and protocols. This would allow
The WBEM technology has the following the integration of distributed management services
features: provided by different management platforms and
applications. Consolidate and unify the data pro-
• Real Standard: Back by many industry, vided by existing management technologies is the
like Microsoft, IBM, Sun, HP, and so on. key purpose of the WBEM initiative.
• Mature implementations exist: Reduce
development and maintenance effort.
316
Table 1. Existing WBEM implement
Producer or Programming Elements and

initiator Platform language License capabilities
CIMOM, Client
Solaris 8 (Spark/
API, Provider API,
Intel) SISSL (Sun
MOF Compiler,
WBEM Services Sun Microsystems open source: plat- Java 1.2 or higher Industry Standards
CIM Workshop,
form-independent Source License)
providers for Solaris
(Java)
OS (only in Solaris)
Linux, Unix
The Open Group CIM Object Broker
(AIX, HPUX,
Pegasus (IBM, Compaq, HP, C++ Open Source (CIMOM), Client
Solaris), Windows
BMC) API, Provider API
NT/2000/9x
Storage Networking
platform-indepen- SNIA Public
SNIA WBEM Industry Association Java (since 1.1.8) Open Source
dent (Java) License
(SNIA)
CIMOM, Client
API, Server API,
Unix, Linux, So- MOF Compiler,
OpenWBEM Caldera C++ Open Source
laris, other WQL (WBEM
Query Language)
tools
C++ (Native Pro- Common Public a set of providers
SBLIM IBM Linux
vider Interface) License for Linux system
API (WBEM
any language Integrated with MS Provider DLL),
MS Windows
WMI Microsoft capable to use DLL Windows operating functionality based
98/2000/XP
API system on the operating
system kernel
General Public libCIM, PaulA
B4WBEM B4WBEM Linux Perl
License (CIMOM API)
simple network management SNMP is based on a manager and agent model.

protocol (snmp) An SNMP-managed network consists of three main
components including: managed devices, agents
The Simple Network Management Protocol and Network-Management Systems (NMSs).
(SNMP) is an application layer protocol that fa- Figure 8 illustrates the relationships of these three
cilitates the exchange of management information components model. These managed objects are
between network devices. Absolutely this protocol arranged in a virtual information database known
is used in the IP-based home network environment. as a Management Information Base (MIB). The
SNMP enables network administrators to manage ‘manager’ is the console where network manage-
IP-based home network performance, find and ment functions are performed, and the ‘agents’
solve network problems. SNMP is a request-reply are the entities or software modules that interface
protocol running over UDP through TCP operation with the actual devices being managed.
is possible. SNMP is extensible, allowing vendors A Management Information Base (MIB) is a
to easily add network management functions to collection of information which defines the prop-
their existing products. SNMP is not a mere paper erties of managed objects. Managed objects are
specification, but an implementation that is widely comprised of one or more object instances, which
available today. are essentially variables. The SNMP manager and
317
Figure 8. NMS, agents and managed devices model for SNMP
agent use a MIB and relatively small set of com- wbem compared to snmp
mands to exchange information of the managed
devices that contain network nodes or managed In this section, WBEM will be compared against
objects. An object identifier or object ID uniquely SNMP. Table 2 shows the different between
identifies a managed object in the MIB hierarchy. WBEM and SNMP. As for the conventional
The MIB hierarchy can be depicted as a tree with network management protocols, the Simple
a nameless root, the levels of which are assigned Network Management Protocol (SNMP) is the
by different organizations. For example of the most popular among the Internet Protocol Based
Figure 9 illustrates the MIB tree. The top-level (IP-BASED) local area network, while no obvious
object IDs represent different organizations, while
the lower-level object IDs are allocated by associ-
ated organizations. The diagram below illustrates Figure 9. Example of the MIB tree with various
an MIB tree where the top-level object IDs are hierarchies
different organizations, and the lower-level ob-
ject IDs are the associated object. To identify the
‘system’ illustrated in the diagram, a unique ID
for the object can be ‘iso.org.dod.internet.mgmt.
mib-2.system’ or the equivalent numeric object
descriptor, ‘1.3.6.1.2.1.1’.
Addition to IP-base network, as PLC networks
and their applications grow with the advances in
PLC technologies, major PLC chipset and modem
vendors are trying to provide network manage-
ment capabilities in their devices by defining
their own private management information base
(MIB). Some PLC management researches (Park
et al. 2008; Park et al. 2008) focus on integrated
management of multi-vendor PLC networks by
using SNMP technology.
318
winner among other types of networks. However, managing archiTecTure

SNMP is not a suitable protocol for remote home and diagnosTic meThod
network diagnosis system because (1) the polling For remoTe conFiguraTion
mechanism used by SNMP may cause network oF heTerogeneous
congestion in a large network or crossing the local neTworks
boundary of wide area network (WAN), and SNMP
places the burden of data collection entirely on the As the communication and network technologies
management side, (2) SNMP agent cannot provide progress rapidly, the number of network-related
the historic record of an equipment or data, and facilities used in the home and small business is
(3) SNMP cannot use a unified data descriptor to also increasing. The raising of the information
preserve the flag, state and configuration of all appliances enables the popularity of easy-to-use,
the managed equipments. Also has follow reasons convenient, and Internet connectable appliances
besides these. The conventional SNMP network in the home and small business networks. In ad-
management approaches have several difficulties dition, the co-existence of different heterogeneous
and challenges (Chen et al., 1998). First is incom- inter-connection technologies, such as IEEE 802.3,
patibility and platform-dependence in managing IEEE 802.11, Bluetooth, IEEE 1394, power line,
heterogeneous networks. Second is high invest- home plug, and so on, is becoming a common
ment. Comparatively, Web-Based management phenomenon.
approaches provide the following advantages: In complicated and heterogeneous network en-
lower cost, scalability, platform-independence, vironment, the management of the home and small
more flexibility and remote access, and easier business network is both a harsh challenge and an
dynamic information. urgent demand for the users, appliance designers
and service providers. In general, when a home
network or a small business network encounters a
Table 2. Comparison between WBEM and SNMP
Features SNMP WBEM

Data encoding format Binary encoding (BER) XML (xml/CIM)
Data description model MIB (Management Information Base) MOF (Managed Object Format)
Namespace Defined by OID (ObjectIDentifiers) Defined by CIM Schema
Data structure Vendor-based CIM-based object model
Transportation protocol UDP HTTP/TCP
Protocol encapsulation IP,UDP, SNMP IP, TCP, (SSL, SHTTP), HTTP, XML
Reliability UDP is connectionless by TCP functions
SNMPv1: Get, GetNext, Set, Trap
SNMPv2: all from SNMPv1, GetBulk, Inform,
Protocol messages types SimpleReq, SimplerRsp, MultiReq, MultiRsp
Reply
SNMPv3: all from SNMPv2, new Trap and Report
HTTP 1.1
SNMPv1, SNMPv2 - only community name SN-
Security SSL+HTTP
MPv3 - USM (User-based Security Model
SHTTP
Management flexibility Low, generally network devices only High, management model is device independent
Other protocol intercon- Mapping to many protocols: SNMP, DMI, CMIP
Low may be realized by proxy agents.
nectivity itd (Mappings are described by DMTF standards)
319
problem, the user usually does not have sufficient common information module (CIM) to seek a
knowledge or expertise to perform diagnosis or solution for each request. The remote LAN mod-
trouble-shooting. Therefore, this service provides ule receives the unsolved requests from the LAN
a good business opportunity of potentially lucra- management module via a channel, configures the
tive revenue for the telecommunication or Internet heterogeneous local networks and uses compat-
service provider (ISP) companies. ible interface at a remote side to accomplish the
As a result of automotive environment is more management and diagnosis for the heterogeneous
complicated and changeable than home environ- local networks.
ment, we take our managing system into home Figure 11 show the module details of the pro-
network, to validate and perform the managing posed architecture. In following section, we detail
functionalities of proposed architecture. The the proposed architecture of remote managing of
proposed managing architecture and diagnostic heterogeneous local area network.
method for remote configuration of heteroge-
neous local networks, which includes at least one sub-network agents
sub-network agent, a local area network (LAN)
management module and a remote LAN module, Each sub-network agent manages the sub-network
as shown in Figure 10. Each sub-network agent through its own management protocol, and collects
manages its sub-networks via its own management sub-network information. The LAN management
protocol, and collects the sub-networks’ informa- module is coupled to the sub-network agent and
tion. The LAN management module receives a cross-internet channel respectively, receives
the requests from heterogeneous local networks one or more requests from the heterogeneous
via these sub-network agents, and converts the local network through the sub-network agent,
information associated with each request into a and converts the information accompanying the
Figure 10. Architecture of remote managing of heterogeneous local area network
320
Figure 11. Module details of the proposed ar-

node information inside the sub-network, infor-
chitecture
mation between nodes, and other related network
information. Each sub-network agent uses its
own managing protocol, such as SNMP or com-
mon management information protocol (CMIP),
to obtain the sub-network information. Each
sub-network build-in, customize or filter out the
analysis of the event relevance. Each sub-network
agent also communicates with LAN management
module through protocol conversion module, and
receives the instruction from LAN management
module to control related sub-networks. The infor-
mation after sub-network agent filters or integrates
is returned to LAN management module so that
remote LAN module knows the information and
the status of each sub-network.
local area network (lan) module
The LAN module provides internet connection

to the remote LAN module through a channel,
and receives one or more management and net-
work information of one or more heterogeneous
local networks through the sub-network agent
and converts into a common information model
for providing the remote LAN module a unified
management information and operation. When the
heterogeneous local network encounters problems,
the LAN management module searches for solu-
request into a common information model to find a
tion first, and when the LAN management module
solution for the request. The remote LAN module
fails to provide solution, the LAN management
receives the requests, which the LAN manage-
module requests the remote LAN module or a
ment module fails to solve, through the channel,
remote manager for assistance so as to realize the
accesses and configures the heterogeneous local
management and diagnosis of the heterogeneous
network remotely, and uses interface compatible
local network.LAN management module includes
with the heterogeneous local network to realize the
a LAN control module, a network information
management and diagnosis of the heterogeneous
module, a network protocol conversion module,
local network.
a common information storage media, a user
Sub-network act as drivers and interface
interface, and a security module.
between the LAN management module and the
Network protocol conversion module is re-
Sub-network. Sub-network agent collect informa-
sponsible for communicating and coordinating
tion of sub-network, and warn about or respond to
with sub-network agent through related protocol
the related event automatically. The information
and mechanism to obtain the basic and man-
provided by the sub-network agent includes the
agement information of sub-network. Network
321
protocol conversion module issue a service, information module. Common information storage
such as through registration, to notify other media is a database or a specific format file, such
related services of the service content so that as management object format (MOF) file matching
other related services use the new service, such CIM in the management infrastructure of WBEM.
as through installation pr-executing instruction. The information stored in common information
Network protocol conversion module converts the storage media is only accessed through network
obtained information into a common information information module.
model or similar format. The common informa- LAN control module is the control center of
tion model is an overall management information LAN management module, providing basic net-
model describing all the computer systems and work management functions, network diagnosis
network equipments in an enterprise network process, algorithm for solving network problems,
environment, including a set of specifications and network information update process, and com-
a set of schemas. Network protocol conversion municating with remote LAN control module.
module transmits the converted information to The basic network management functions include
network information module. The data conversion numerous items such as allowing the user to know
is such as converting the sub-network instruction through user interface the network internal basic
into common information model. The functions information, such as network topology, network
of network protocol conversion module could be traffic, network speed, and node information, or
added flexibly. even the software installation on each node. In the
Network information module analyzes and problem diagnosis, the collected information is
organizes the converted common information used to analyze the possible cause of the problem,
from network protocol conversion module, and and forwards the problem to LAN control module.
then selects and defines a common information If LAN control module fails to solve the problem
model suitable for storing network information internally, a request is sent to remote LAN module
for storing in common information storage media. for assistance to provide a solution.
Network information module is responsible for ac- When the problem is solved, LAN control
cessing information stored in common information module or remote LAN control module report to
storage media required by LAN control module the user or warn the user for preventing similar
for diagnosis and management, and receives events in the future. When the problem is not
the issued diagnosis and managed instructions completely solved, LAN control module or remote
and passes the instructions to network protocol LAN control module will also inform the user of
conversion module for performing diagnosis and the problem handling status. For example, through
management. The common information model has the node analysis to obtain the node equipment
sufficient expressive capability to represent all the information, LAN control module or remote LAN
managed objects, and has sufficient expansion control module use the information to search for
capability to accommodate new managed objects equipment manufacturer for repairing and post
as well as accessing the management information the information on the user interface to inform the
effectively. For example, the common information user. The cause or the solution to the problem is
model object manager (CIMOM) of the man- recorded on common information storage media
agement infrastructure of web-based enterprise or remote common information storage media for
management (WBEM) plays the role of network future reference.
information module. User interface show the network information,
Common information storage media is the ac- such as network topology, each node information,
tual information storage media to match network network traffic, and network speed. Through the
322
graphic and web-based interface showing the cally operate in a way similar to those modules
network status, the user easily use the mouse or in LAN management module. The difference lies
button to process complicated network problem in remote LAN control module is designed for
and issue management instructions with simple remote LAN diagnosis and management service
and clear guiding mode. For the user, the inter- center. Therefore, remote control module moni-
face is simple and effective. Therefore, the user tor, manage, and register LAN, such as all home
understands the network internal basic information network and small business network. In general,
through the use of the user interface. remote LAN module collects and utilizes global
Security module is responsible for related se- data, such as periodically obtaining the topology
curity mechanism, such as security authentication of each LAN. In performing LAN management
mechanism, data encryption mechanism, protect- and diagnosis, remote LAN control module will
ing the internal data access mechanism for LAN request LAN control module of a certain LAN
control module, and billing mechanism. Security management module for detailed information, and
authentication mechanism is to verify whether stores the information in remote information stor-
the user is the legitimate user or administrator so age media for assisting the solving of the problem
as to protect LAN control module from invasion that LAN control module of LAN management
by illegitimate remote LAN control module. Dur- module fails to solve.
ing information transmission, a data encryption
mechanism is provided to prevent data from theft. algorithms
The data encryption mechanism is controlled by
security module and remote security module, A diagnosis method for remote configuration
respectively. These two security modules both of heterogeneous local networks comprises the
encrypt and decrypt the data transmitted and steps of: detecting through at least a sub-network
received on either side. The encrypted data will agent whether any sun-network connecting to the
not be easily theft or utilized so as to achieve the sub-network agent encounters any problem; when
security objective. detecting at least a sub-network encountering
problem, collecting the management and network
remote lan management module information of the sub-network with the problem,
and transmitting to a LAN management module;
Remote LAN management module receives the converting the management and network infor-
unsolved request from LAN management module mation of the sub-network with the problem into
through channel, such as request from sub-net- a common information model through the LAN
work, remotely access and configures the hetero- management module, and performing diagnosis
geneous local network, and adopts an interface to determine whether a solution can be provided;
compatible to the heterogeneous local network when the LAN management module fails to pro-
to realize the management and the diagnosis of vide a solution, requesting a remote LAN module
the heterogeneous local network. Remote LAN or a remote management for assistance to realize
module includes a remote LAN control module, the management and diagnosis of the sub-network
a remote network information module, a remote with the problem; and responding the diagnosis
common information storage media, a remote user message to a user interface or the sub-network
interface, and a remote security module. agent. Figure 12 illustrates diagnosis methods
Remote LAN control module, remote network for remote configuration of heterogeneous local
information module and remote common informa- networks.
tion storage media of remote LAN module basi-
323
A method for updating network informa- store remotely, informing a remote LAN module
tion for a management architecture for remote and storing the updated information to a remote
configuration of heterogeneous local network, common information storage; and if not to store
comprising the steps of: determining whether remotely, storing the updated information to a local
network information is periodic information common information storage. Figure 13 illustrates
or dynamic information; if periodic, using the network information update process.
extract instruction to obtain the update informa- A method for adding a sub-network to a
tion; if dynamic, collecting and processing the managing architecture for remote configuration
network information and converting the processed of heterogeneous local networks, comprising the
network information into a common information steps of detecting a new network through a net-
and obtaining the update information; filtering work protocol conversion module; determining
the updated periodic or dynamic information the attributes of the new network through a LAN
and determining whether to store remotely; if to management module; when the LAN management
Figure 12. Diagnosis methods for remote configuration of heterogeneous local networks
324
Figure 13. Network information update process
module fails to provide the support service to the Vehicular sysTem

new network, sending a request to a remote LAN archiTecTure
module; the remote LAN module finding a suit-
able sub-network agent to the new network and Figure 15 show the system architecture. The
transmitting to the LAN management module; and system architecture contains two parts. The left
when the new network having a sub-network agent, part is OSGi platform and the right part is WBEM
performing the registration and announcement of server. The OSGi platform contains some bundle
the new network. Figure 14 illustrates process of build upon OSGi framework; include authentica-
adding a new sub-network. tion bundle, vehicular network diagnosis bundle,
IP network agent bundle, Power Line network
agent bundle, IEEE 1394 network agent bundle
and connect interface bundle. The bundles further
describe as follow:
325
Figure 14. Process of adding a new sub-net-

forwards the problem to vehicular network
work
diagnosis bundle. If vehicular network di-
agnosis bundle fails to solve the problem
internally, a request is sent to remote man-
ager for assistance to provide a solution.
• Connect interface bundle: Connect inter-
face bundle is adapter between OSGi plat-
form and WBEM server. All information
collected by network agent will convert to
CIM model by connect interface bundle
and store on WBEM repository or pass to
remote center. Connect interface bundle
also receive the instance from WBEM
server.
• Authentication bundle: Authentication
bundle is to verify whether the user is the
legitimate user or administrator so as to
protect home server from invasion by ille-
gitimate remote client.
The remote client provides a user interface

that manager could watch the vehicular network
information, manage device or sent diagnosis
information to local home server if local don’t
have enough solution.
wbem-based remote Vehicular

network management and
diagnosis system
• Network agent bundle: The IP network Figure 16 shows the WBEM-based Remote
agent bundle collects information of IP Vehicular Network Management and Diagnosis
network and uses SNMP managing proto- System GUI. The WBEM-based Remote Vehicu-
col to manage device. Other network agent lar Network Management and Diagnosis System
do same thing but using its own managing GUI contains three parts: the left panel shows the
protocol. vehicular network topology by using tree view;
• Vehicular network diagnosis bundle: the top-right panel contains some buttons, and
Vehicular network diagnosis bundle pro- the bottom-right panel, which display messages
viding basic network management func- about the detail device information and allow sent
tions, network diagnosis process, algorithm solutions info to WBEM-based Local Vehicular
for solving network problems. In the prob- Network Management and Diagnosis System.
lem diagnosis, the collected information The vehicular network topology is local vehicu-
by network agent bundle is used to ana- lar network topology include IEEE 1394 network
lyze the possible cause of the problem, and topology, IP-base network topology, and Lon-
326
Figure 15. System architecture
Figure 16. WBEM-based remote vehicular network management and diagnosis system GUI
Works network topology. When click the device All process in WBEM-based Remote Vehicular
on left-panel topology, top-right panel will show Network Management and Diagnosis System are
the detail information of this device. For example matching management object format (MOF) file
in Figure 16, by clicking the Port7 device we could CIM in the management infrastructure of WBEM.
get the device description is Hardware x86, device The MOF of system includes attribute and method
uptime is 1 day 21:29:20.5, location is restroom, of system process. The lonworksdeviceinfo,
IP address is 192.168.1.50 and some other detail ieee1394deviceinfo, and ipdeviceinfo are the
information show on top-right panel. defined network information model. Information
327
Figure 17. XML for invoking/return the snmpX method
of each sub-network is described in this model. value by input ipaddress and oid. The findip is the
The topologyX is the method operation that client method that do mac to ip convert. In addition, the
could get local sub-network topology. The snmpX new CIM model could add to repository through
is the method operation that client could get oid- OSGi update or download.
Figure 18. WBEM-based Local Vehicular Network Management and Diagnosis System GUI
328
Figure 19. Diagnosing of the changes of SNMP switch network topology
All information show on remote client is through the listen port of transport mappings to
through WBEM server and is CIM model. It uses get the Ethernet network information.
RMI or CIM Operations over HTTP to transmit In the LonWorks network topology adopts the
information between the residential gateway and web page to simulate the i.Lon 100 web page. The
remote manager. The operations and messages are i.Lon 100 web service can control the devices on
described abstractly, and then bound to a concrete and off or set up the time when the devices will be
network protocol and message format to define an on and off, and it also shows the device status.
endpoint. The CIM Operations over HTTP using When the topology changes, the text panel will
an XML format for describing network services. show the possibility problems to let user fixed the
Figure 17 illustrates invoked and returned opera- problem easier. Figure 19 illustrates diagnosing
tions for transmitting managing message. process and the changes of SNMP switch network
topology after plug in the IP-based device. That
wbem-based local Vehicular SNMP network topology is changed from three
network management and nodes into the four nodes.
diagnosis system All process on WBEM-based Local Vehicular
Network Management and Diagnosis System are
The WBEM-based Local Vehicular Network matching management object format (MOF) file
Management and Diagnosis System main GUI and CIM in the management infrastructure of WBEM.
OSCAR bundle list is shown in Figure 18. The No matter is topology information or diagnosis
GUI contains three parts: the top-left panel shows knowledge.
the vehicular network topology by using tree view;
the top-right panel contains some buttons, and the
bottom panel, which display messages about the reFerences
home network situation or receive information
from remote client. Ai, Y., Sun, Y., Huang, W., & Qiao, X. (2007).
The client GUI shows IEEE1394 network topol- OSGi based integrated service platform for au-
ogy, SNMP switch network topology, LonWorks tomotive telematics. In Proceedings of the IEEE
network topology. In the IEEE1394 network topol- International Conference on Vehicular Electronic
ogy that adopts the driver of the libraw1394 to get and Safety, (pp. 1-6).
the network information because Java language Chen, J. (1998). A Study of Web-Based SNMP
doesn’t support the IEEE1394 drive. Network Management with a Simple Java Applet
In the SNMP switch network topology adopts Network Monitoring Tool. Department of Com-
the SNMP4J API that classes are capable of cre- puter Science and Engineering Auburn University,
ating, sending, and receiving SNMPv1/v2c/v3 Alabama.
messages. SNMP4J receives SNMP messages
329
Kim, J. H., Seo, S. H., Moon, T. Y., Hwang, S. H., Park, J. G., Ahn, C. W., Cho, H. N., Byun, I. S.,
& Jeon, J. W. (2007). A method of improving the Desmons, F., & Kim, S. W. (2006). A method for
reliability of the gateway system by using OSEK/ representing and transporting CIM operations
VDX. In Proceedings of IEEE International using binary XML in the WBEM architecture. In
Conference on Control, Automation and System, Proceedings of the 8th International Conference
(pp. 2328-2833). on Advanced Communication Technology, (Vol.
3, pp.20-22).
Mariño, P., Poza, F., Dominguez, M. A., & Otero,
S. (2009). Electronics in automotive engineering: Saito, T., Tomoda, I., Takabatake, Y., Arni, J., &
A top-down approach for implementing industrial Teramoto, K. (2000). Home gateway architec-
field bus technologies in city buses and coaches. ture and its implementation. IEEE Transactions
IEEE Transactions on Industrial Electronics, on Consumer Electronics, 46(4), 1161–1166.
589–600. doi:10.1109/TIE.2008.2002723 doi:10.1109/30.920474
Marples, D., & Kriens, P. (2001). The Open Ser- Sun, Y., Huang, W. L., Tang, S. M., Qiao, X., &
vices Gateway Initiative: An Introduction Over- Wang, F. Y. (2007). Design of an OSEK/VDX
view. IEEE Communications Magazine, 39(12), and OSGi-based embedded software platform for
110–114. doi:10.1109/35.968820 vehicular applications. In Proceedings of the IEEE
International Conference on Vehicular Electronic
Park, C. K., Kang, J. M., Choi, M. J., Hong, J. W.,
and Safety, (pp. 1-6).
Lim, Y. H., & Choi, M. S. (2008, Jan.). An Inte-
grated Network Management System for Multi- Thompson, J. P. (1998). Web-based enterprise man-
Vendor Power Line Communication Networks. In agement architecture. IEEE Communications Mag-
Proceedings of the IEEE International Conference azine, 36(3), 80–86. doi:10.1109/35.663331
on Information Networking, (pp.23-25).
Park, C. K., Kang, J. M., Choi, M. J., Hong, J. W.,
Lim, Y. H., Ju, S., & Choi, M. S. (2008). Defini-
tion of common PLC MIB and design of MIB
Mapper for multi-vendor PLC network manage-
ment. In Proceedings of the IEEE International
Symposium on Power Line Communications and
Its Applications, (pp.152-157).
330
331
Chapter 19
Using Wireless Mesh Network
for Traffic Control
Kun-Chan Lan
absTracT
Wireless mesh networks (WMN) have attracted considerable interest in recent years as a convenient,
flexible and low-cost alternative to wired communication infrastructures in many contexts. However, the
great majority of research on metropolitan-scale WMN has been centered around maximization of avail-
able bandwidth, suitable for non-real-time applications such as Internet access for the general public.
On the other hand, the suitability of WMN for missioncritical infrastructure applications remains by and
large unknown, as protocols typically employed in WMN are, for the most part, not designed for real-
time communications. In this chapter, we describe a real-world testbed, which sets a goal of designing
a wireless mesh network architecture to solve the communication needs of the traffic control system in
Sydney, Australia. This system, known as SCATS (Sydney Coordinated Adaptive Traffic System) and used
in over 100 cities around the world, connects a hierarchy of several thousand devices -- from individual
traffic light controllers to regional computers and the central Traffic Management Centre (TMC) - and
places stringent requirements on the reliability and latency of the data exchanges. We discuss some
issues in the deployment of this testbed consisting of 7 mesh nodes placed at intersections with traffic
lights, and show some results from the testbed measurements.
inTroducTion the environment via a reduction in fuel consump-

tion. One of the main and most common functions
Adaptive traffic control systems are employed in of such systems lies in adaptive control of traffic
cities worldwide to improve the efficiency of traf- lights. This ranges from simple lengthening or
fic flows, reduce average travel times and benefit shortening of green and red light durations in an
intersection according to the actual presence of cars
DOI: 10.4018/978-1-60566-840-6.ch019 in the respective lanes, to coordination of green light
Using Wireless Mesh Network for Traffic Control
phases among neighboring intersections on main from them, in the anticipation that their major use
thoroughfares. This adaptivity is made possible will be public, for purposes such as accessing the
with the use of sensors (typically in the form of Internet or conducting voice calls (Ganguly et al.,
magnetic loop detectors embedded under the road 2006). On the other hand, little attention has been
pavement) that feed data to roadside traffic light directed to the aspects of reliability and latency,
controllers, and a communications infrastructure which are most important if MAWMN are to be
that connects among the intersections and a traf- considered for replacement of mission-critical
fic management centre, as well as, in some cases infrastructure, such as traffic control system
(typically in large cities), a hierarchy of regional communications.
computers (RC) that perform the control decisions In this chapter, we describe a wireless mesh
for respective portions of the system. network testbed that has been built. The testbed
Traditionally, the communications layer of physically covers seven traffic lights in the sub-
traffic control systems has been based on wired urban area of Sydney. These intersections are
connections, either private or leased from public chosen because they represent a typical suburban
telecommunications operators. While for many area with lots of traffic, foliages, pedestrians and
years such leased lines (operating at 300bps) high-rise residential buildings. In addition, the
have served their purpose well, they have several inter-node distance (ranging from 200 to-500m)
shortcomings, such as a significant operating is representative of 90% of the distance between
cost, inflexibility, and difficulty of installation in traffic controllers in the Sydney CBD (Central
new sites. In certain cases, alternative solutions, Business District) area. The testbed nodes have
operating over public infrastructure, have been been custom-built.
deployed for specific sites where private or leased The contribution of this paper is three-fold.
lines were not a viable option; these ranged from First, to the best of our knowledge, our work is
ADSL, regular dialup, or cellular (GPRS). How- one of the first efforts to study the feasibility of
ever, using public network for traffic control could using wireless mesh networking for traffic control.
suffer from inconsistent delay jitters and reliability Second, we describe the details of our testbed
issues. For example, previous experimental studies implementation and some experiences we gained
(Chakravorty et al., 2002) have shown GPRS links during the deployment of the testbed in an urban
could have very high RTTs (>1000ms), fluctuating environment. Finally, we present some initial
bandwidths and occasional link outages. measurement studies of link characteristics of
In recent years, there has been considerable different wireless and wired technologies used in
interest in wireless mesh networks and their our testbed (including the use of 900MHz, 2.4GHz
deployment in metropolitan areas, from both a and 3.5GHz radios and Ethernet-over-powerline).
commercial and a research perspective (Lundgren Although our results are still very preliminary,
et al., 2006). Trials in several major cities in the they are useful to serve as a reality check toward
US (e.g. Philadelphia, New Orleans, and others the goal of applying wireless mesh networking
(Tropos networks, 2009; Locust world, 2009) and to traffic control applications.
worldwide (e.g. Taiwan (Mobile Taiwan) have The rest of this paper is structured as follows.
shown mesh networks to be a viable technology In section 2, we describe the details of SCATS, the
that can compete well with alternative “last-mile” traffic control system used in Sydney and many
connectivity solutions to the public. Correspond- other cities worldwide, and its communication
ingly, most of the research on metropolitan-area requirements. We describe related work in Sec-
wireless mesh networks (MAWMN) has focused tion 3. Section 4 presents a simple analysis of
on maximising the throughput that can be extracted the topology of traffic lights in the Sydney met-
332
ropolitan area, and in particular the dependence detectors, are predominantly made using serial
of the degree of connectivity of the mesh on the point-to-point connections over standard voice-
radio range. Section 5 describes the details of our grade telephone lines, using 300bps modems. This
testbed implementation. We present some initial is also the most common method of connecting be-
measurement results of link characteristics of tween the TMC and other lowbandwidth devices,
different radio technologies used in our testbed including variable message signs, variable speed
in section 6. Section 7 discusses the experiences limits, ramp meters, and over-height detectors.
we gained during the deployment of our testbed. At the core of the SCATS operation is a periodic
We conclude the paper and discuss the future exchange of messages between the controlling
work in section 8. computer and each and every intersection (via
the point-to-point links). This exchange happens
every 1sec, and is initiated by the computer which
scaTs oVer wireless sends to the intersection’s local controller a com-
mand message, instructing it about the next phase
In this section, we first describe the details of it should switch to and the timing of that switch.
SCATS (Sydney Coordinated Adaptive Traffic The controller, in turn, is required to reply with
System) and its communication requirements. an acknowledgement, which includes information
We then discuss the benefits and research chal- from the intersection’s sensors. If an acknowledge-
lenges when running SCATS on a wireless mesh ment is not received within 1sec from the time
network. the command message is sent, it is retried once;
after the second time an acknowledgement fails
The scaTs Traffic to arrive, the communications link is declared
management system failed, and SCATS instructs all controllers at the
respective cluster of intersections to fall back into
Developed and maintained by the Roads and a ‘default’ self-controlling mode, where decisions
Traffic Authority (RTA, formerly Department of about the timing of green light phases are made
Main Roads) of the state of New South Wales, locally and independently. Likewise, a controller
the Sydney Coordinated Adaptive Traffic System will fall back to this mode upon not receiving a
(SCATS) is one of the most popular traffic manage- command message. Once triggered, a controller
ment systems used worldwide. Its main task is to will stay in the self-controlling mode for at least
adjust, in real time, signal timings in response to 15 minutes; if another communications failure
variations in traffic demand and system capacity. happens during this time, the duration of this
Real-time data from traffic controllers are collected mode will be extended by another 15 minutes,
and transported to a central traffic management and so on. Obviously, the self-controlling mode,
centre (TMC) for analysis and optimum control where the decisions at intersections are uncoor-
of road traffic. The performance of SCATS, there- dinated, can lead to a severely suboptimal traffic
fore, depends critically on the capabilities of the control, particularly in a busy thoroughfare during
underlying communication system that transports rush hour. Accordingly, though the bandwidth
roadside data to and from the TMC. required from the communication links is quite
The existing communication system of SCATS low (comfortably handled by 300bps modems),
relies strongly on third-party wired infrastructure the 1sec latency is critical for an efficient opera-
(provided by Telstra, Australia’s largest telco). The tion of the system.
bulk of the communications to the intersections,
namely the traffic light controllers and vehicle
333
The currently used SCATS infrastructure, in wireless technology provide bandwidth that is
based on wired communications, suffers from the more than adequate for connecting many high-
following problems: resolution roadside cameras to SCATS.
One possible option for going wireless is to
• Slow installation and inflexibility. In most build a dedicated RTA wireless network using
cases, installing a new line at a road site widely available, standards based, low-cost wire-
(especially a remote site) involves earth less technologies, e.g. IEEE 802.11x and 802.16x.
excavation, which is very slow and with 802.11x equipment is cheaper, less complex, and
adverse effects on existing infrastructure. operates entirely in the unlicensed spectrum (no
• High capital and operating cost. The in- licensing fee). On the other hand, 802.16x is more
stallation of a wired connection at a new reliable (has multiple carrier frequencies to avoid
site, or repairs at an existing one, carries interference), has longer range, and better features
a high cost due to the material and labour to cater for a diverse range of communication
required. More importantly, the ongoing needs of future roadside equipment. In addition,
fees for leasing the wires from the tele- it is possible to operate 802.16x in both license
phone company run very high; currently, and unlicensed spectrums.
RTA pays nearly $40 million annually to Despite of its enormous benefits, there are
Telstra in leasing fees for connecting the several challenges when roadside ITS equipment
traffic signals and other roadside devices to is connected via wireless media:
SCATS.
• Low bandwidth. Modem-based leased • Latency. Wireless can potentially increase
lines support bandwidth less than 32 Kbps. latency. For example, IEEE 802.11x, uses
While these low-bandwidth telephone lines a common wireless channel (it is cheaper
are adequate for connecting traffic signal to share channel) among many contend-
sensors, they cannot provide adequate sup- ing devices causing potential conflict. To
port for connecting high-bandwidth appli- avoid such conflicts, some form of medium
cations, e.g. higher solution video cameras, access control (MAC) is implemented by
which increasingly becoming necessary to these technologies. MAC introduces some
effectively monitor traffic pattern on our delay before data can be transmitted on the
roads. wireless channel.
• Reliability. Wireless signals are suscep-
going wireless tible to interference from other signals in
the vicinity operating in the same or adja-
With wireless solutions, there is no cabling cent spectrum. Given that ITS equipment
involved. Wireless can therefore provide fast in- is deployed in public area, such interfer-
stallation and exceptional flexibility. Cost can be ence will be the norm rather than excep-
reduced significantly by building a private wire- tion. Interference can corrupt messages
less network, because there will be no monthly transmitted over the wireless medium.
charges to be paid to telephone company (some Some frequencies do not work well (or at
small license fee may apply). Moreover, the in- all) if there is no direct line-of-sight be-
stallation cost will be low because there will be tween the two communicating end points.
no cabling-related labour. The cost issue is, in In a dynamic context of public roads,
fact, the major concern for most road authorities. roadside equipment may frequently face
Finally, it should be noted that recent advances line-of-sight problems due to transient
334
obstructions, e.g. a high vehicle carrying a described the impact of distance, SNR and trans-
tall crane etc. Also in vehicle-to-roadside mission rate on the packet loss. While Roofnet’s
communications, a car in the near-lane propagation environment is characterized by
may obstruct communication between a its strong Line-of-Sight component, our work
far-lane car and the roadside equipment. differs from the prior work in that our links are
Temporary outages, i.e., periods when no generally heavily obstructed 1. In addition, our
wireless signal is available, therefore, is a planned deployment strategy is different from the
real issue to deal with. unplanned topology in Roofnet.
• Security. What makes wireless so vulner- For example, our antenna is mounted at a height
able is the fact that the attacker does not of about 4 meters from the ground. But the trees
have to gain physical access to the channel on the road are typically higher than 7 meters.
from any predefined access point. Roadside Similar to our work, The WAND project
wireless components are well within the (Weber et al., 2003) has built a multi-hop wire-
wireless range of passing motorists and less testbed in the centre of Dublin. They have
pedestrians, which make them vulnerable 11 nodes mounted on traffic lights along a 2km
to intrusion, denial of service, and other route in urban area. However, their topology is
forms of security threats. simpler than ours (i.e. a chain topology) and the
• Scalability. As mentioned earlier, wireless measurements they performed on their testbed
systems are sensitive to interference from were relatively limited.
other communicating devices operating TFA project (Camp et al., 2006) aimed to pro-
in the vicinity. Additionally, if a common vide broadband access to low income community
wireless channel is shared among all de- in Houston area via wireless mesh network tech-
vices within a given area (cell), the MAC nology. Their architecture consist of two wireless
delay increases rapidly as the number of tiers: an access tier to connect homes, businesses,
competing devices increases. Another and mobile users to the infrastructure, and a back-
scalability issue arises from the process- haul tier to forward traffic to and from the wired
ing overhead that is required at a central entry point. Jardosh et al. (2005) discussed the
radio base-station. The more remote radios correlation of link reliability with the frame re-
there are in communication with the cen- transmissions, frame sizes and data rate by collect-
tral radio, the more processing that must ing trace data from a structured 802.11b network
take place. The radio controller at the base- during a international conference. They concluded
station will simply not be able to process that sending smaller frames and using higher data
all incoming radio signals if there are too rates with a fewer number of frames improves the
many of them. performance of congested network.
All the previous studies have been centered
around maximization of available bandwidth for
relaTed work non-real-time applications such as broadband
access for the general public. On the other hand,
Roofnet (Bicket et al., 2005) is an experimental this testbed is the first to focus on using wireless
802.11b/g mesh network built by MIT. Each node mesh networking for traffic control which places
in Roofnet has an antenna installed on the roof stringent requirements on the reliability and la-
of a building. Aguayo et al. (2004) analyzed the tency of the data exchanges.
link layer behavior on the Roofnet testbed and
335
a simple analysis oF that range) between every pair of traffic controllers,

The sydney TraFFic and the output was then sorted so that for each
lighT Topology controller, its neighbours were listed (from nearest
to furthest) with the distance to each neighbour. We
This analysis was based on data provided by RTA, then processed this data to determine how many
indicating the latitude and longitude of traffic neighbours of each traffic controller were within
controllers. Figure 1 shows points at each traffic a specified range (from 100m to 1250m).
controller location. As shown in Figure 1, there The results of this analysis are shown in Figure
are around70 controllers in a 4km × 6km area. 3, which shows on the y-axis how many traffic
To understand the effect of radio range on the controllers had 0, 1, 2, 3, . . . neighbours when a
degree of connectivity when the traffic controllers given radio range is assumed (x-axis). The results
are forming a mesh network, we calculate the shown in Figure 3 provide a rough indication of
shortest distance (assuming that the radio has a what radio range is needed if we are aiming to
circular radio range and have perfect coverage in interconnect a certain number of nodes with each
node having a certain degree (number of neigh-
bours within range). For example, if we seek to
Figure 1.Location of traffic controllers interconnect 90% of the nodes (accepting that
some alternative technology may be needed for
the minority 10% of nodes) and require that each
node has three neighbours (to provide redundancy
and hence fault tolerance), then we require a radio
technology with range of at least 1km. Note that,
while city environments may have large densities
of traffic controllers in both (lat/long) dimensions,
in suburban environments controllers (particularly
those that are important to synchronise with com-
munication links) often lie linearly along main
arterial roads, with few controllers in close range
Figure 2. Map of intersection locations
336
orthogonal to the main arteries. Neighbours that at intersection m1 is connected to a gateway node
form a line would not provide the same level of in University of Sydney.
fault tolerance as those that are better separated The streets where the network is deployed are
angularly around a controller, since the links are about 10- 20m wide and surrounded by building
less likely to fail independently. at least two stories high. The majority of these
buildings are made of concrete and steel that
block the propagation of radio signals into the
TesTbed neighboring streets. All these streets have busy
public traffic during business hours. Most of the
In this section, we provide the details of this tes- vehicles on the street have a height of less than
tbed. We first describe the environment that the 2.5m. But some double-decker buses (such as
testbed is located. Next, the hardware used and Sydney Explorer Bus) or truck can have a height
the software installed on each of these nodes are of more than 5m.
discussed.
channel characteristics
environment
Wireless channels can be coarsely characterized
The testbed is located in the Sydney CBD (Central by its path loss exponent. Pathloss described the
Business District) area. Seven intersections are attenuation experienced by a wireless signal as a
selected to deploy the testbed, as shown in Figure function of distance. The amount of variations in
2 (specifically, intersection number m1 to m7). pathloss between similar propagation scenarios
A number of custom-build embedded PCs with is called shadowing. In other words, shadowing
multiple wireless interfaces are used. The nodes represents the difference between the signal power
are mounted on the traffic lights at a height of at different points in the same environment with
about 2-3m from the ground, and distributed along the same estimated pathloss. Prior study (Rapport
the streets in the form of rectangle covering an et al., 1996) showed that shadowing can be mod-
area of 500 × 1000 square metres at a distance of eled as a zero-mean Gaussian random variable.
200-500m apart. None of the nodes is in a clear Specifically, one can predict the received signal
line of sights of its neighboring nodes. The node power at a given distance d with the following
formula:PdBm(d ) = PdBm(d 0) - 10a log 10(d
d 0) + _ where a is the pathloss exponent, _ is
Figure 3.Numbers of neighbours within certain the shadowing and d 0 is the reference distance.
radio range of RTA controllers The prior work (Rapport et al., 1996) suggested
that the pathloss can range from 2 to 5 for outdoor
urban environment.
Such physical level measurements are impor-
tant for an efficient deployment (i.e. overestimat-
ing pathloss can result in overprovioning network
while underestimating pathloss can produce dis-
connected network). By using linear regression,
the environment of this testbed is found to have
a pathloss a = 3.1 and shadowing _ = 7.2 .
Note that the observed pathloss in this environ-
ment is significantly lower than the suggested
337
urban pathloss of 4 in the literature (Rapport et • Wireless interfaces: Each node has two
al., 1996). wireless interfaces to connect to its neigh-
boring nodes, as shown in Figure 5. To al-
hardware low the testbed users to experiment with
different radio technologies, two different
The hardware components used for the nodes radio frequencies are currently used on the
in this testbed are off-the-shelf products includ- testbed: 2.4GHz (802.11b/g) and 900MHz
ing the following, as shown in Figure 4. All the radios. Specifically, the nodes at intersec-
components are mounted on two sides of a metal tion m2, m3 and m6 are installed with
plate for easy maintenance (one can simply swap two 2.4GHz mini-PCI wireless cards from
an old plate with a new plate when we want to Ubiquiti (SR2). The nodes at intersections
upgrade the node). A custom-built enclosure is m1 and m5 are equipped with one 2.4GHz
made to house this component plate. Ubiquiti SR2 card (with a transmission
power of 400mW) and one 900MHz
• Motherboard: A VIA MB720F Mini-ITX Ubiquiti SR9 card (with a transmission
motherboard featuring an 1GHz processor power of 700mW). Finally, the nodes at
and 1GHz memory is employed as the core intersections m4 and m7 are installed with
in our system. two Ubiquiti SR2 cards. One of these two
• Storage: The traffic pole sometimes vi- SR2 cards is connected to a 2.4GHz-to-
brates a lot due to the passing traffic. Since 900MHz converter (from RFlinx) to send
that our node is mounted on a traffic pole, 2.4GHz signal output by the wireless card
instead of using a hard-drive, we employ on 900MHz band. Due to its better pen-
a 2G USB flash drive for storing OS and etration rate for buildings and trees, theo-
data. Unlike a hard-drive, a flash drive retically the use of 900MHz radios could
does not have a high speed spinning platter result in a better connectivity than 2.4GHz
and is less failure-prone. radios (i.e. 802.11x). Hence, 900MHz
Figure 4. Hardware component
338
radios are used at intersection pairs m1- link to remotely connect the mesh node in
m5 and m4-m7. These two intersection addition to Unwired and Ubiquiti links.
pairs have a longer distance (i.e. 400m and • Power: As shown in Figure 4, an off-the-
500m respectively) than the other intersec- shelf power-board (with surge protector
tion pairs. and fuse) and a PC power-supply are used
• Back-haul connection: In addition to the to provide the power to all the compo-
two Ubiquiti wireless cards, each node is nents in the node. The power-board takes a
equipped an “Unwired” modem (Unwired 240AC power from the traffic light.
wireless, 2009) to establish a back-haul • Antenna: Nodes on this testbed are all in-
link back to NICTA for the purpose of re- stalled with omni-directional antennas due
mote management, as shown in Figure 5. to the following
Unwired is a Sydney-based metropolitan • Cost: An omni-directional antenna is typi-
wireless ISP. The Unwired modem uses a cally cheaper than a directional antenna. In
proprietary protocol but claims to be a vari- addition, for a node which has n neighbors,
ant of WiMAX and operates in a licensed n directional antennas are needed. On the
3.5GHz band. other hand, one omni-directional antenna
• Ethernet router: A Linux-based Ethernet per intersection is sufficient to cover all the
router (Diamond Digital R100) is installed neighbors.
in each node. We employ this Ethernet • Mounting. The space on the traffic light
router for several purposes. First, it is used for the mounting of antennas is quite lim-
as an Ethernet switch to connect the moth- ited. It is comparatively more difficult to
erboard and the Unwired modem (and any mount a directional antenna on the traffic
IP-based devices such as a camera in the pole in practice. An 8dBi omni-directional
future). The Unwired modem is connect- antenna is used for the 2.4GHz wireless
ed to the WAN port of the router, thus the card and an 6dBi omni-directional antenna
router get a public Internet IP address. The is used for the 900MHz wireless card.
motherboard has an Ethernet connection • Weatherproof: The temperature in the
with the router’s 4-port switch. Second, the summer can be above 40 Celsius degree in
Diamond Digital router has an USB port Sydney. The temperature inside the node
which allows the motherboard to have a se- can be even higher. As shown in Figure 4,
rial connection with the router’s USB port to provide enough air circulation inside the
through an USB-to-serial adapter. Being node, many holes are drilled on the bottom
able to establish a serial link to the moth- of the enclosure and made some air louvres
erboard allows the user to remotely login on the side. Two temperature-controlled
into the serial console for troubleshooting fans are used in the node to dissipate the
when the Ubiquiti wireless interfaces are hot air out through the louvres. In addition,
not responding. Third, given that the router a roof is mounted on top of the enclosure
is a Linux box itself (runs on openWRT), to shield the enclosure from direct sunlight
we can run all the existing software (e.g. and rain.
we are currently running DNS, NTP and • Remote recovery: Due to the fact that the
VPN clients on it). Finally, the Diamond testbed is deployed in an outdoor envi-
Digital router has an 802.11 wireless inter- ronment, it is time consuming to visit the
face which can be used as an alternative nodes when something goes wrong. In ad-
dition, given that nodes are mounted on the
339
Figure 5. Testbed topology
traffic lights which are a public asset, visit- encapsulate the SCATS data from the serial port
ing any node on the testbed required calling of the traffic controller board into an IP packet
out the RTA maintenance crew to gain ac- as well as to encapsulate the IP packet from the
cess to the node. Therefore, some means of regional computer and send its payload to the
remote recovery are necessary. Therefore, serial interface.
one wireless remote switch is installed on In order to connect the testbed to the regional
each node (runs in the unlicensed 433MHz computer, a gateway node at University of Sydney
band), which allows one to reboot the node is deployed. The gateway node has a reasonable
on-site when accessing the node via the line-of-sight to intersection m1 and connects to the
2.4GHz or 3.5GHz links fails. m1 node with a 802.11 link. Note that the Unwired
links are not used to connect the regional computer
The ultimate goal of this testbed is to control (RC) to the testbed due to the consideration of reli-
traffic lights using wireless mesh networks. A ability, latency and cost issues. More details about
pair of power-over- Ethernet adapters (Netcomm the characteristics of Unwired links are described
NP285) is used to connect the node to a traffic in Section 6. The RC is connected to the gateway
controller board in the curbside housing through node via AARNet (Aarnet - australia’s research and
the powerline. The traffic controller board sends education network, 2009). The round-trip delay
and receives data via a serial interface. Hence, a between RC and the gateway is about 1.2ms, and
serial-to-IP conversion is required for the com- the throughput is typically over 100Mbps.
munication between the traffic controller and the
testbed (which runs IP). The traffic controller software
board is mounted inside an embedded PC and
connected to the embedded PC’s motherboard’s The testbed node uses a custom-built Linux OS im-
(VIA MB770F) serial port. A serial-to-IP con- age that consists of the following components:
verter software is written and run on the PC to
340
• The image size is small enough to be fit link latency

into an USB flash drive. And run com-
pletely in RAM (1GB). This allows one to The round-trip delay increases as the number of
enable ’clean’ reboots uncontaminated by hops increases on the 802.11 links. In addition, the
previous experiments. variation also increases significantly when there
• Some kernel modifications are added for are more hops. There is no strong correlation be-
various driver supports for USB, madwifi tween distance and link latency though. As shown
and PXE reboot. in Figure 6, the latency does not increase from
• Grub is used and modified to activate the 300m to 400m. However, the variation increase
watchdog timer at the time of boot-loading significantly as the distance increases. It is due to
so that the watchdog timer can be started that there are more retries at 300m than at 400m
before any program start. Watchdog timer due to different line-of-sight conditions.
is used to reboot the motherboard when the We next examine the efficiency of power line
system fails. communication. As suggested in Figure 7, given
• Various tools are used including time- a distance of 100m, the link latency of power
sync, Open- VPN, some network sup- line communication is excellent. The average
port tools and software from Orbit project round-delay is about 3.6ms and the variations
(Raychaudhuri et al., 2005) in our image. are very small. In addition, the largest delay for
The image is built to be Debian-based for such a distance is less than 8ms. As described in
the compatibility with Orbit software. Section 5, the Unwired network is used to carry
out our back-haul traffic. To understand the ex-
The OS image is based on DSL-N (Damn small pected latency of running management traffic
linux). DSL-N provides most of the software sup- over the Unwired network, the round-trip delay
port we need out of the box. The default syslinux to the mesh node is measured. As shown in Figure
bootloader of DSL-N is replaced with grub. OML 8(a), the average delay of sending traffic over
software (Singh et al., 2005) from Orbit project the Unwired network to the mesh node is about
is used to build the measurement collection infra- 400ms. However, there is a large variation (the
structure for the testbed. Two security mechanisms delay can be as long as 3 seconds) and significant
are currently implemented. First, OpenVPN is used number of outages. The delay and outages over
for the Unwired links to each of the mesh nodes.
Second, ssh and friends are used on all network
interfaces. In addition, root access is disabled on Figure 6. Effect of distance on round-trip delay
all the machines.
link characTerisTics
In this section, some results of measured link

characteristics from the testbed are described.
Specifically, some statistics of the wireless link
performance are discussed in terms of round-trip
delay, loss and throughput. Ping is used to measure
the round-trip delay and iperf for the throughput
measurement.
341
Figure 7. Latency of powerline communication

However, the loss becomes more bursty as the
number of hops and distance increase. Note that
Figure 9 and Figure 10 are based on the results
from a low probing rate (i.e. one-packet-persecond
ping). The loss pattern might change if we change
the probing rate. In addition, we do not find there
is a strong correlation between packet loss and the
distance. The line-of-sight condition (which is
location-dependent) plays a more important role on
the packet loss. The use of 900MHz radio results in
a much lower loss rate (0.5%) than 2.4GHz radio
(20%), which is not surprising though since
900MHz radio has a better penetration rate than
2.4GHz radio. Finally, for a distance of 100m, the
loss rate of power line communication is almost
the Unwired network are mostly contributed by
negligible (less than 0.1%).
the wireless link between the mesh node and the
Unwired ase station. As shown in Figure 8(b), the Throughput
average delay of the Unwired wireless link is about
200ms. The large delay variations and significant As shown in Figure 11, the use of 900MHz radio
number of outages suggest that a public-shared results in a better throughput than when 2.4GHz
wireless network like Unwired is not suitable for radio is employed, even for a longer distance.
operating SCATS traffic. There is a larger variation when using 900MHz
radio, which might be due to MAC-layer retrans-
losses mission. The throughput of Ethernet-over-power
line communication is very stable and typically
As shown in Figure 9 and Figure 10, the packet
maintained at about 600Kps.
loss seems to be distributed uniformly over time.
Figure 8. Round-trip delay over the Unwired network
342
Figure 9. Effect of number of hops on consecutive Figure 10. Effect of distance on consecutive
packet loss packet loss
Figure 11. Comparison of throughput for different technologies
deploymenT issues is transmitted into the air. Some protection

of the antenna plug might be necessary be
In this section, we discuss some issues in terms of necessary for an operational network to
the deployment and maintenance of our testbed ensure there is no signal leakage from the
in an urban environment. antenna connector. In addition, while the
appearance of the hardware might look
deployment identical, it is safer to check if the hard-
ware does comply with the specification
• Hardware. Many antenna connectors were before starting using it. For example, some
held on by weak glue or crimp. Gradual of Senao wireless cards used do not output
stress (e.g. vibration) could eventually loos- a transmission power of 200mW as they
en the plug and degrade the signal before it should be according to the specification.
343
• Software. Most of the wireless measure- the upper antenna might be less obstruct-
ments are based on readings from the ed and hence has a better connectivity. At
wireless cards. However, while the hard- 2.4GHz, a quarter wavelength is approxi-
ware can be identical, different firmwares mately 30cm. Antenna position changes
and drivers could introduce inaccuracy in in the range 10-30 cm can cause dramatic
the measurement results. If possible, one changes in signal strength, presumably due
should try to validate the readings from to the presence of standing waves in the vi-
a wireless card against the results from a cinity of the traffic light pole or more spe-
spectrum analyser. cifically in the vicinity of metal stop signs
• Antenna locations. As described in and the like! If multiple antennas are de-
Section 5, each node is equipped with ployed, it is essential to have a means for
three antennas, including two 2.4GHz independently adjusting their position.
(or one 2.4GHz and one 900MHz) omni-
directional antennas and one 3.5GHz di-
rectional antenna. To facilitate the ease of maintenance
mounting, all three antennas are mounted
on a pole and then mount this pole on the • Remote management: Remote manage-
traffic light. Specifically, one omni-direc- ment is challenging, the following ways
tional antenna is pointing upward and the are provided to allow the user to access the
other is pointing downward while the di- nodes.
rectional antenna is mounted in between. • To access the Linux-based Ethernet
The location of antenna can have an effect router: The router can be connected via the
on the link performance. Figure 12 shows Unwired link over OpenVPN. In the case
the round-trip delays from node m2 to when the OpenVPN connection can not be
its neighboring node m3 using the omni- established, given that a public IP address
directional antennas. The use of the lower is obtained for each router from Unwired,
antenna results in a higher loss and a larger one can connect to the router via its public
variation. One possible explanation is that IP address, although this could introduce a
Figure 12. Effect of antenna locations on round-trip delay
344
dependency on a dynamic DNS lookup. In • Security: Security is a major concern es-

addition, one can connect to the mesh node pecially when the wireless mesh testbed is
first and then connect to the router via the sharing public spectrum with an average
Ethernet or USB-to-serial link between the of 50+ external APs at each intersection.
router and the motherboard. Furthermore, the testbed is effectively con-
• To access the mesh node: One can connect nected to the Internet via Unwired network,
to the mesh node (i.e. the motherboard) by and exposed to various password attacks.
first connecting to the router and then con- In a live deployment for traffic control, the
nect to the motherboard via the Ethernet or mesh security should be integrated with
USB-to-serial link. Being able to access the the traffic control system security model,
motherboard via its serial port is important which may include e.g. segmentation to
since the Ethernet link might fail for vari- contain the damage of a denial of service
ous reasons. In addition, one can access the or break-in attack, combined with multiple
motherboard via its 802.11 interfaces from levels of fallback to local control.
any reachable neighboring nodes. In addi-
tion, the following mechanisms are imple- conclusion
mented to recover the system from failure.
First, and the default way to recover from a In this chapter, we discuss some issues in deploying
failure is to login to the offending router or a testbed as a first step towards creating a fully
motherboard using one of the above meth- functional wireless mesh network-based traffic
ods, analyse the problem and/or reboot the control system. In addition, we describe some
node. Second, the watchdog timer support initial results of link characteristics of different
on the MB720 is used. In addition, Grub is technologies used on our testbed. Many research
setup to fall back to a stable backup image challenges such as latency, reliability, security and
which is installed in a separate partition in scalability are need to be addressed.
case when the default image fails. Finally,
the BIOS is configured to give top priority
to PXE network boot but DHCP server is reFerences
configured in a way that it does not provide
PXE boot information in the default case. Aarnet - australia’s research and education net-
Therefore the node defaults to its second work. (2009). Retrieved from http://www.aarnet.
priority, which is to boot from the USB edu.au/
flash drive. However, in the event of a node Aguayo, D., Bicket, J., Biswas, S., Judd, G., &
failure (for example, due to a bad image) Morris, R. (2004). Link-level measurements from
the DHCP server can be quickly reconfig- an 802.11b.
ured to support the PXE boot. Having re-
booted the node using PXE, a new working Bicket, J., Aguayo, D., Biswas, S., & Morris, R.
disk image can be distributed to the node (2005). Architecture and evaluation of an un-
via frisbee or FTP. In practice, one might planned 802.11bmesh network. In Proceedings
use FTP instead of frisbee since that frisbee of the 11th annual international conference on
introduce more control traffic overhead on Mobile computing and networking(MOBICOM),
the Unwired links, which charges for every Cologne, Germany. Available from http://www.
bits send to the nodes. pdos.lcs.mit.edu/papers/roofnet: mobicom05/
roofnet-mobicom05.pdf
345
Camp, J., Robinson, J., Steger, S., & Knightly, Mobile taiwan applications promotion project
E. (2006). Measurement driven deployment of a m-taiwan. Retrieved from http://www.pwlan.org.
two-tier urban mesh access network. In Proceed- tw/mp.asp?mp=3
ings of ACM MobiSys 2006, Uppsala, Sweden.
Rapport, T. S. (1996). Wireless communications
Available from http://networks.rice.edu/papers/
principles and practice. Upper Saddle River, NJ:
sys7122-camp.pdf
Prentice Hall.
Chakravorty, R., & Pratt, I. (2002). Performance
Raychaudhuri, D., Seskar, I., Ott, M., Ganu, S.,
issueswith general packet radio service. Journal of
Ramachandran, K., Kremo, H., et al. (2005).
Communications and Networks (JCN), 4(2). Avail-
Overview of the orbit radio grid testbed for
able from http://www.cl.cam.ac.uk/Research/
evaluation of next-generation wireless network
SRG/netos/papers/comob-web/2002-jcn.pdf
protocols. In Proceedings of the IEEE Wireless
Damn small linux not (dsl-n), (n.d.). Retrieved Communications and Networking Conference,
from http://www.damnsmalllinux.org/dsl-n/ New Orleans, LA, USA.
Ganguly, S., Navda, V., Kim, K., Kashyap, K., Singh, M., Ott, M., Seskar, I., & Kamat, P. (2005).
Niculescu, D., Izmailov, R., et al. (2006). Perfor- Orbit measurements framework and library (oml):
mance optimizations for deploying voip services in Motivations, design,implementation, and features.
mesh networks. Retrieved from http://www.wings. In Proceedings of IEEE Tridentcom 2005, Trento,
cs.sunysb.edu/%7Eanand/papers/jsac06.pdf Italy.
Jardosh, A., Ramachandran, K., Almeroth, K. C., Tropos networks. (2009). Retrieved from http://
& Belding-Royer, E. M. (2005). Understanding www.tropos.com
congestion in ieee 802.11b wireless networks. In
Unwired wireless. (2009). Retrieved from http://
Proceeding of ACM SIGCOMM Internet Mea-
www.unwired.com.au/
surement Conference, Berkeley, CA. Retrieved
from http://moment.cs.ucsb.edu/amitj/jardosh- Weber, S., Cahill, V., Clarke, S., & Haahr, M.
imc2005.pdf (2003). Wireless ad hoc network for dublin: A
large-scale ad hoc network test-bed. ERCIM
Locust world. (2009). Retrieved from http://www.
News, 54. Retrieved from http://www.ercim.org/
locustworld.com
publication/ErcimNews/enw54/weber.html
Lundgren, H., Ramachandran, K., Belding-Royer,
E., Almeroth, K., Benny, M., Hewatt, A., Touma,
A., & Jardosh, (2006). Experiences from building
and using the ucsb meshnet testbed. IEEE Wire-
less network. Available from http://user.it.uu.se/
[UNKNOWN ENTITY &tildent;]henrikl/publica-
tions.html mesh network.
346
Section 7
Mobility Model, Simulation,
and Security
348
Chapter 20
Mobility Models of
Vehicular Networks
Kun-Chan Lan
absTracT
A key component for VANET simulations is a realistic vehicular mobility model that ensures conclusions
drawn from simulation experiments will carry through to real deployments. However, VANET simulations
raise many new questions about suitable levels of details in simulation models. To get accurate results,
the mobility models of Vehicular Networks should be as realistic as possible, and involve road-maps
with all constraints and facilities related to the vehicular movement. In this chapter, the authors provide
an overview of some mobility models that are relevant to VANETs. The criteria of applicability they
consider here is the employment of road maps, and thus limiting the nodes movements into the routes,
instead of moving them in a wide open area. They compare different models based on the parameters
they use. For instance, some models use traffic control mechanisms (stop signs or traffic lights) at route
intersections, and some just assume continuous movement at these points. Some assume routes to be
single-lane, some others support multi-lanes routes. Some define the security distance, while others just
ignore this parameter.
inTroducTion VANETs. One of the realistic applications of mobile

Ad hoc Networking (MANET) is Vehicular Ad hoc
To get accurate results, the model should be as Networking (VANET). In VANET, communica-
realistic as possible, and involve road-maps with tions between nodes do not rely on any dedicated
all constraints and facilities related to the vehicular infrastructure. Although it is self-organized and easy
movement. In this chapter, we provide an over- to deploy, the infrastructureless vehicular network
view of some mobility models that are relevant to introduces many challenges that should be tackled
before real implementations. For instance, to allow
DOI: 10.4018/978-1-60566-840-6.ch020 communications between nodes (vehicles) which are
Mobility Models of Vehicular Networks
out of the power range of each other, some other on the performance results. However, except city
intermediaries should act as routers to remedy the section (Zheng et al., 2004), all the other models
lack of dedicated routers. Thus, a distributed rout- presented in this survey are inappropriate for
ing protocol needs to be employed. It is mandatory VANETs. In (Bettstetter et al., 2001; Zheng et al.,
before passing to the real deployment of a routing 2004) the authors classify the models according
protocol as well as any other protocol or applica- to the randomness of speeds and directions, and
tion) to evaluate it by simulation. The faithfulness divide them into i) trace-based (deterministic)
of the simulation results depends on the realism models, ii) constrained topology based models
of the parameters and the models used in the and iii) statistical (fully random) models. Based on
simulation, particularly on the mobility model. this classification, (Zheng et al., 2004) provides a
This can be defined as the pattern that establish more recent survey of MANET’s mobility models,
the nodes movement within simulation area during presenting some ones that consider obstacles in
a simulation. The earlier mobility models used the simulation area. These models are suitable to
in MANET simulation assume the terrain to be simulate pedestrian movements, but not vehicles.
without obstacles, and nodes to be able to move In (Boudec et al., 2005), the authors provide a
freely in the whole rectangular stimulation area. mathematical and simulation investigations into
For example, Random way-point (RWP) (David some of the statistical (random) models, and
et al., 1996) is a typical example of such a kind combine them in the so-called random trip. The
of models, which is largely used in the literature, most recent research works on mobility models
and available in many network simulators (ns2, focus on the vehicular ad hoc network application
GloMoSim, etc.). This model defines the pause- of MANET. (Luo et al., 2004) provides a general
time parameter, so that each node has phases of survey of VANET, and the existing challenges to
movement and others of pause. At the beginning, overcome before the real deployment. In (Chisalita
the node selects randomly and uniformly a desti- et al., 2004) the authors discuss the usefulness of
nation where it goes using a random and constant VANET to ensure the vehicular traffic safety and
speed, which is also selected for each movement facilities, as well as the advantages it provides
following a uniform distribution. Once it reaches compared to the other centralized technologies.
this destination it stays there for the pause-time Some specified applications have been proposed,
duration, then repeats the process. It has been il- such as the discovery of free parking places
lustrated that this model engenders, after a given (Caliskan et al., 2006). Regarding the mobility
simulation time, a spatial distribution of nodes models, some new ones have been specifically
concentrated around the center of the simulation proposed for VANET, such as (Amit Kumar Saha
area (Bettstetter et al., 2001 ; Bettstetter et al., et al., 2004 ; Choffnes et al., 2006 ; Mahajan et al.,
2003). Generally speaking, the assumption of an 2006; Gorgorin et al., 2006; Karnadi et al., 2007).
open terrain is realistic for some applications of In this manuscript, we first present and discuss
pedestrians, but it is inappropriate for VANET. these novel models, as well as those proposed for
More recent studies propose new models con- the general MANET that apply to VANET. We
strained to routes and obstacles, and thus are more also present a new vehicular traffic simulator we
suitable for VANET. In (Camp et al., 2002) Camp implemented to generate movement trace files
et al. discuss a variety of mobility models used to usable by some well-known network simulator,
evaluate ad hoc networks, and split them up into notably GloMoSim (Zeng et al., 1998) and ns2
two categories: entity models and group models. (The Network Simulator, 2009), then we use this
The authors show by simulation how the choice tool along with GloMoSim to conduct a simulation
of a mobility model can have a significant effect comparative study. Our main contribution here is
349
the investigation into the overtaking impact on manhaTTan

the network performance.
In the following we provide an overview on This is also a generated-map-based model,
the models that were either specially devoted to introduced in (Bai et al., 2003) to simulate an
VANETs, or proposed in the general MANET urban environment. Before starting a simula-
context but usable in VANET. The criteria of ap- tion, a map containing vertical and horizontal
plicability we consider here is the employment roads is generated. Each of these latter includes
of road maps, and thus limiting the nodes move- two lanes, allowing the movement in two direc-
ments into the routes, instead of moving them in tions (north/south for the vertical roads and east/
a wide open area. As we will see, the considered west for the horizontal ones). At the beginning
parameters differ from one model to another. of a simulation, vehicles are randomly put on
For instance, some models use traffic control the roads. Afterwards, they move continuously
mechanisms (stop signs or traffic lights) at route according to history-based speeds (exactly like
intersections, and some just assume continuous Freeway). When reaching a crossroads, the vehicle
movement at these points. Some assume routes randomly chooses a direction to follow. That is,
to be single-lane, some others support multi-lanes continuing straightforward, turning left, or turn-
routes. Some define the security distance, while ing right. The probability of each decision is set
others just ignore this parameter. by the authors respectively to 0.5, 0.25, 0.25. The
Freeway is a generated-map-based model, security distance is also used in this model, and
defined in (Bai et al., 2003). The simulation area, nodes follow the same strategy as in the freeway
represented by a generated map, includes many model to keep this distance. But contrary to the
freeways, each side of which is composed of many previous model, a vehicle can change a lane at
lanes. No urban routes, thus no intersections are a crossroads. Nonetheless, there is no control
considered in this model. At the beginning of the mechanism at these points (crossroads), where
simulation, the nodes are randomly placed in the nodes continue their movements without stopping,
lanes, and move using history-based speeds. The which is unrealistic.
speed of each vehicle smoothly changes following
a random acceleration. In addition to the realism
related to the acceleration and the history based ciTy secTion mobiliTy (csm)
speed, the model defines a security distance that
should be maintained between two subsequent CSM (Zheng et al., 2004) can be viewed as a
vehicles in a lane. If the distance between two hybrid model between RWP and Manhattan, as
vehicles is less than this required minimal distance, it introduces the principle of RWP, especially
the second one decelerates (its acceleration is the pause-time and random selection destination,
forced to be negative) and let the forward vehicle within a generated-map-based area. At each step of
moves away. The change of lanes is not allowed the vehicle’s movement a random point is selected
in this model. The vehicle moves in the lane it from the generated road map, toward which it
is placed in until reaching the simulation area moves following the shortest path. After reaching
limit, then it is placed again randomly in another that destination, it remains there for a pause-time,
position and it repeats the process. This scenario then repeats the process. The speeds of nodes are
is definitely unrealistic. constrained by the security distance, along with
the maximum speed limit of the road.
350
real map model (rmm) with this model is the unrealistic disposition of the
stop signals, since it is impossible to find a region
Thus far, we have presented models based on with stop signals at each intersection. Therefore,
virtual generated maps. RMM (Amit Kumar they proposed TSM (Mahajan et al., 2006), which
Saha et al., 2004) is very similar to CSM, but we describe hereafter.
indeed it uses real maps, obtained from the TI-
GER/Lines database (Amit Kumar Saha et al.,
2004). For each route segment 1, the coordinates TraFFic sign model (Tsm)
are extracted and converted using the Mercator
projection (Amit Kumar Saha et al., 2004). The In this model, stop signals are replaced by traf-
extracted points are then presented by a graph, fic lights. A vehicle stops at a crossroads if it
where the crossroads are presented by vertices, encounters a red stoplight, otherwise it continues
and routes by weighted arcs. The weight of each its movement. When the first vehicle reaches the
arc is dynamically calculated in such a way to intersection, the light is randomly turned red, with
mimic the estimated time required for a vehicle probability p (thus turned green with probability
to move over the corresponding segment, which is 1 − p). If it turns red, it remains so for a random
proportional to its maximum authorized speed, its delay (pause-time), forcing the vehicle to stop as
distance, and the number of vehicles it currently well as the ones behind it. After the delay, it turns
contains. Therefore, the lower the weight, more green then the nodes traverse the crossroads one
the vehicles move freely in the segment. Note that after the other until the queue is empty. When the
the maximum authorized speed of a route segment next vehicle arrives at the crossroads the process
depends on its type. Finally, we mention that like is repeated. TSM and SSM have been evaluated
all the previous models, RUM defines no control by simulation with ns2 (The Network Simula-
mechanisms at crossroads. tor, 2009). The results show that the two models
are not significantly influenced by the speed of
nodes (maximum speeds). This is due to the traf-
sTop sign model (ssm) fic control models, which slow down the nodes
and give more stability to the network (Mahajan
Contrary to the previous models, SSM (Mahajan et al., 2006). When increasing the pause-time at
et al., 2006) integrates a traffic control mechanism. the intersections, the authors remarked that the
In every crossroads, a stop signal is put, which performances improved for both models, and that
obliges vehicles to slow down and make a pause SSM gives better results than TSM when using
there. This model is based on real maps of the the same pause-time. The authors argue this by
TIGER/Lines database, but all roads are assigned the fact that in SSM nodes always stop at the in-
a single lane in each direction. A vehicle should tersections, unlike TSM. Nevertheless, in reality
never overtake its successor (like in all the models the pause-time for stop signals is shorter than that
presented before), and should tune its speed to keep of traffic lights, which makes TSM more stable
the security distance. If many vehicles arrive at an indeed (Mahajan et al., 2006).
intersection at the same time, they make a queue,
and each one waits for its successor to traverse the
crossroads. This results in gathering of nodes, and sTraw TraFFic
hugely affects the network connectivity as well as
the vehicle mobility (average speeds). According STRAW (Choffnes et al., 2006) is also a model
to the authors (Mahajan et al., 2006), the problem relying on real maps of TIGER/Line. Like the
351
other models (except freeway), the roads include moVe

one lane in each direction, and is divided into
segments. The model is basically composed of MOVE (Karnadi et al., 2007) is a VANET’s
three modules: intra-segment mobility manager, mobility model that uses as the compiler SUMO
inter-segment mobility manager, and finally the (SUMO Simulation of Urban Mobility, 2009),
route management and execution module. At the which is a realistic vehicular traffic simulation
beginning of the simulation, the nodes are placed model. SUMO is an open-source application
randomly one behind the other. Then they move implemented with java, that integrates many re-
using the car following model (Choffnes et al., alistic parameters, such as realistic accelerations,
2006) such that they try to accelerate until reach- the usage of real maps reflecting several types
ing the maximum speed of the segment. The first of routes (with multiple lanes), as well as traffic
module manages this movement until reaching an lights defining priorities between vehicles. Basi-
intersection. The security distance is maintained, cally, MOVE is composed of two components:
and the overtaking is not allowed. At crossroads the road map editor and the vehicle movement
the vehicles always slow down, even when they editor. The former serves to manually and ran-
change a segment and turn without a full stop, domly generate a road map, either from TIGER/
which is realistic. The second module defines the line files or Google earth files, whereas the latter
traffic control mechanism, which includes both allows to specify the properties of each vehicle,
stop signals and traffic lights, put on crossroads like the maximum speed, the acceleration, the
according to the class of the intersected routes. probability of turning at crossroads, the path to
In addition to this usual control form, the module take etc. The information collected by the two
makes sure that the next segment to take contains editors is sent to the SUMO compiler, then a trace
enough available space before moving the vehicle file in ns-2 or Qualnet format is generated. MOVE
toward it. If it is fully busy, the vehicle waits at has been compared by simulation to RWP using
the crossroads (at the end of the first segment). AODV. The results show that MOVE causes a
The last module selects the routes to be taken by low reception rate.
each vehicle during the simulation. It implements
two approaches: simple straw and straw OD. In
the first one, the direction is randomly selected gorgorin eT al. model
at each intersection. That is, when reaching an
intersection, the vehicle randomly decides whether In addition to all the realistic parameters of the
to continue straightforward, or to turn and change previous models, this one (Gorgorin et al., 2006)
the route. On the other hand, in the second ap- implements an overtaking mechanism within
proach a destination is selected toward which multi-lane segments. A vehicle always tries to
the vehicle moves using the shortest path. The move on the most right lane (the lowest rang),
simulation study made by the authors (Choffnes except in case of overtaking during which it moves
et al., 2006) show that when using STRAW the left, and intersections in urban environments
reception ratio decreases from 43% up to 53% where it chooses the lane according to the next
compared to movements in an open area. The direction. A hierarchy of vehicle states is defined,
results of this simulation also illustrate that the respectively: free driving, approaching, follow-
routes arrangement has an impact. Scenarios ing, and braking (in the order). When a vehicle
with a high number of crossroads slow down the is in another state than the free driving, it checks
average speeds of nodes, which improves the whether higher lanes allow it to pass to a higher
reception ratio. state, and thus moves to the left lane to make an
352
Table 1.
Manhat-
Feature\ Model Freeway tan CSM RUM SSM TSM STRAW MOVE Gorgorin
Real maps no no no yes yes yes yes yes yes
# lanes/direction many one one one one one one many many
Intersections no yes yes yes yes yes yes yes yes
Changing lanes
no yes yes yes yes yes yes yes yes
at intersections
traffic
traffic
Traffic control no no no no stop signs both lights+ both
lights
priority
Overtaking no no no no no no no no yes
Security dis-
yes Yes yes no yes yes yes yes yes
tance
Pause-time no No yes no yes yes yes yes yes
overtaking. Identically, a vehicle in a state different Bai, F., Sadagopan, N., & Helmy, A. (2003). Im-
than braking checks whether the right lane allows portant: a framework to systematically analyze
it to at least stay in the same state and then moves the impact of mobility on performance of routing
right. Moreover, the model allows to specify the protocols for ad hoc networks. In The 22th IEEE
driver type, which affect many parameters of the Annual Joint Conference on Computer Commu-
vehicles (speed, acceleration, etc.). Finally, note nications and Networking INFOCOM’03, (pp.
that the model includes both traffic lights and stop 825–835).
signals at intersections. One of these two different
Bettstetter, C. (2001). Smooth is better than
control mechanisms is put at each intersections
sharp: a random mobility model for simulation
according to the types of the intersecting seg-
of wireless networks. In Proceedings of the
ments. The most important parameter added in this
4th ACM international workshop on Modeling,
model is the overtaking mechanism. However, no
analysis and simulation of wireless and mobile
study investigating this issue has been done yet.
systems (MSWIM ’01) (pp. 19–27). New York:
Table 1 summarizes the features of all the models
ACM Press.
presented in this section.
Bettstetter, C., Resta, G., & Santi, P. (2003). The
node distribution of the random waypoint mobility
reFerences model for wireless ad hoc networks. IEEE Trans-
actions on Mobile Computing, 2(3), 257–269.
Amit Kumar Saha, D. B. J. (2004). Modeling doi:10.1109/TMC.2003.1233531
mobility for vehicular ad-hoc networks. In the 1st
ACM international workshop on Vehicular ad hoc Boudec, J.-Y. L., & Vojnovic, M. (2005). Perfect
networks (pp. 91–92). New York: ACM Press. simulation and stationarity of a class of mobility
models. In The 24th IEEE Annual Joint Conference
on Computer Communications and Networking
INFOCOM’05, (pp. 2743–2754).
353
Caliskan, M., Graupner, D., & Mauve, M. (2006). Johansson, P., Larsson, T., & Hedman, N. (1999).
Decentralized discovery of free parking places. In Scenario-based performance analysis of rout-
Proceedings of the 3rd International ACM Work- ing protocols for mobile ad-hoc networks. In
shop on Vehicular ad hoc networks, VANET’06, MOBICOM 99, The Fifth Annual ACM/IEEE
(pp. 30–39). New York: ACM Press. International Conference on Mobile Computing
and Networking, Seattle, WA (pp. 15–19).
Camp, T., Boleng, J., & Davies, V. (2002). A
survey of mobility models for ad hoc network Karnadi, F. K., Mo, Z. H., & Lan, K. C. (2007).
research. Wireless Communications and Mobile Rapid generation of realistic mobility models for
Computing (WCMC): Special issue on Mobile Ad vanet. In IEEE Wireless Communications and
Hoc Networking: Research . Trends and Applica- Networking Conference.
tions, 2(5), 483–502.
Luo, J., & Hubaux, J.-P. (2004). A survey of inter-
Chisalita, I., & Shahmehri, E. (2004). Vehicular vehicle communication. School of computer and
communication: A candidate technology for traffic Communication Sciences, EPEL, Tech. Rep.
safety. In IEEE International Conference on Sys- IC/2004/24.
tems, Man and Cybernetics, (pp. 3903–3908).
Mahajan, A., Potnis, N., Gopalan, K., & Wang, A.-
Choffnes, D. R., & Bustamante, F. E. (2006). I. A. (2006). Urban mobility models for vanets. In
An integrated mobility and traffic model for Proceedings of the 2nd IEEE International Work-
vehicular wireless networks. In The 2nd ACM shop on Next Generation Wireless Networks.
international workshop on Vehicular ad hoc
SUMO Simulation of Urban Mobility. (2009).
networks, VANET’05, (pp. 69–78). New York:
Retrieved from http://sumo.sourceforge.net/
ACM Press.
The Network Simulator. ns2.(2009). Retrieved
David, B., & David, A. (1996). Dynamic source
from http://www.isi.edu/nsnam/ns/
routing in ad hoc wireless networks. In Mobile
Computing, (vol. 35, pp. 153–181). Kluwer Zeng, X., Bagrodia, R., & Gerla, M. (1998).
Academic. Glomosim: A library for the parallel simula-
tion of large-scale wireless networks. In The
Davies, V. (2000). Evaluating mobility mod-
12th Workshop on Parallel and distributed
els within an ad hoc network. Master’s thesis,
Simulation. PADS’98, Banff, Alberta, Canada
Colorado School of Mines, Colorado, USA,
(pp. 154–161).
Tech. Rep.
Zheng, Q., Hong, X., & Ray, S. (2004). Recent
Gorgorin, C., Gradinescu, V., Diaconescu, R.,
advances in mobility modeling for mobile ad hoc
Cristea, V., & Ifode, L. (2006). An integrated
network research. In Proceedings of the 42nd ACM
vehicular and network simulator for vehicular
annual Southeast regional conference (ACM-SE),
ad-hoc networks. In the 20th European Simulation
(70-75). New York: ACM Press.
and Modelling Conference.
354
355
Chapter 21
MOVE:
A Practical Simulator for
Mobility Model in VANET
Kun-Chan Lan
absTracT
Vehicular Ad-Hoc Network (VANET) is surging in popularity, in which vehicles constitute the mobile
nodes in the network. Due to the prohibitive cost of deploying and implementing such a system in real
world, most research in VANET relies on simulations for evaluation. A key component for VANET simula-
tions is a realistic vehicular mobility model that ensures conclusions drawn from simulation experiments
will carry through to real deployments. However, VANET simulations raise many new questions about
suitable levels of details in simulation models for nodes mobility. In VANET simulations, the mobility
models used affect strongly the simulation output. The researchers need to decide what level of details
are required for their simulations. In this chapter, the authors introduce a tool MOVE that allows us-
ers to rapidly generate realistic mobility models for VANET simulations. MOVE is built on top of an
open source micro-traffic simulator SUMO. The output of MOVE is a realistic mobility model and can
be immediately used by popular network simulators such as ns-2 and Qualnet. The authors show that
the simulation results obtained when using a realistic mobility model such as MOVE are significantly
different from results based on the commonly used random waypoint model. In addition, the authors
evaluate the effects of details of mobility models in three case studies of VANET simulations (specifically,
the existence of traffic lights, driver route choice and car overtaking behavior) and show that selecting
sufficient level of details in the simulation is critical for VANET protocol design.
DOI: 10.4018/978-1-60566-840-6.ch021
MOVE
inTroducTion at the intersection (e.g. turn left, turn right or go

straight). Such a turning model could have an
VEHICULAR Ad-Hoc Network (VANET) com- effect on the congestion of the road as well as
munication has recently become an increasingly on the clustering of the vehicles. Furthermore,
popular research topic in the area of wireless a smooth deceleration and acceleration model
networking as well as the automotive industries. should be considered since vehicles do not abruptly
While it is crucial to test and evaluate protocol break and move. Some prior studies (Saha et al.,
implementations in a real world environment, 2004; Heidemann et al., 2001) have shown that
simulations are still commonly used as a first step a realistic model is critical for accurate network
in the protocol development for VANET research. simulation results. Selecting appropriate level
Several communication networking simulation of details in the mobility model for a VANET
tools already existed to provide a platform to test simulation is a critical decision. Unrealistic
and evaluate network protocols, such as ns-2 (T. N. mobility model can produce simulations that are
S. ns 2, 2009), OPNET (O. Simulator, 2009) and misleading or incorrect. On the other hand, adding
Qualnet (Q. N. Simulator, 2009). However, these details requires time to implement and debug. In
tools are designed to provide generic simulation addition, it might increase simulation complexity,
scenarios without being particularly tailored for slow down simulation, and distract the research
applications in the transportation environment. problem at hand.
On the other hand, in the transportation arena, In this chapter, we develop a tool MOVE (MO-
simulations have also played an important role. bility model generator for VEhicular networks)
A variety of simulation tools such as PARAMICS to facilitate users to rapidly generate realistic
(P. M. T. Simulation, 2009), CORSIM (CORSIM, mobility models for VANET simulations. MOVE
2009) and VISSIM (P. simulation VISSIM, 2009) provides an environment that allows the user to
etc have been developed to analyze transportation quickly pinpoint incorrect details and manage
scenarios at the micro- and macro-scale levels. details overhead. Our tool is built on top of an
However, there was little effort in integrating com- open source micro-traffic simulator SUMO (S. S.
munication techniques and scenarios in a realistic of Urban Mobility, 2009). The output of MOVE
transportation simulation environment. is a mobility trace file that contains information
One of the most important parameters in of realistic vehicle movements which can be im-
simulating ad-hoc networks is the node mobility. mediately used by popular simulation tools such
It is important to use a realistic mobility model so as ns-2 or qualnet. MOVE allows users to rapidly
that results from the simulation correctly reflect generate realistic VANET mobility models in two
the real-world performance of a VANET. A real- aspects:- by interfacing with real world map da-
istic mobility model should consist of a realistic tabases such as TIGER (Topologically Integrated
topological map which reflects different densities GEographic Encoding and Referencing) (TIGER,
of roads and different categories of streets with 2009) and Google Earth, MOVE allows the user
various speed limits. Another important parameter to conveniently incorporate realistic road maps
should be modeled is the obstacles. In the real into the simulation. In addition, by providing a
world, a vehicle node is typically constrained to set of Graphical User Interfaces that automate
streets which are separated by buildings, trees or the simulation script generation, MOVE allows
other objects. Such obstructions often increase the user to quickly generate realistic simulation
the average distance between nodes as compared scenarios without the hassle of writing simulation
to that in an open-field environment. In addition, scripts as well as learning about the internal details
each vehicle needs to decide a turning directions
356
MOVE
Figure 1. The architecture of MOVE

trips of vehicles and the route that each vehicle
will take for one particular trip. We currently sup-
port three different methods to define the vehicle
movements– the vehicle movement patterns can
be manually created by the user, generated auto-
matically, or specified based on a bus time table
to simulate the movements of public transport.
The information users input in the Map Editor
and the Vehicle Movement Editor is then fed into
SUMO to generate a mobility trace which can
be immediately used by a simulation tool such
as ns-2 or qualnet to simulate realistic vehicle
of the simulator. The architecture of MOVE is movements.
shown in Figure 1. Users can also visualize the generated mobility
In this chapter, we first describe the archi- trace by clicking on the “Visualization” button
tecture implementation of MOVE (Section 2). on the main menu, as shown in Figure 3. This is
We then compare MOVE against the commonly useful for observing the details of vehicle move-
used random waypoint model and show that a ment (e.g. cars’ overtaking behavior) and debug-
realistic mobility model is critical for VANET ging (e.g. to avoid incorrect setting of simulation
simulations. We present three case studies that parameters).
consider three different scenarios including the
existence of traffic light, driver route choice at
the intersection, and car overtaking behavior. We
discuss how these details affect the network topol- Figure 2. Mobility model generator
ogy and resultingly the performance of VANET
in the simulation (Section 3).
archiTecTure
MOVE is currently implemented in Java and

runs atop an open-source micro-traffic simulator
SUMO. MOVE consists of two main compo-
nents: the Map Editor and the Vehicle Movement
Editor, as shown in Figure 2. The Map Editor is
used to create the road topology. Currently our
implementation provides three different ways to
create the road map – the map can be manually
created by the user, generated automatically, or
imported from existing real world maps such as
publicly available TIGER database from U.S.
Census Bureau (TIGER, 2009). The Vehicle
Movement Editor allows the user to specify the
357
MOVE
Figure 3. Visualization of vehicle movements

map editor
In MOVE, the road map can be generated manu-

ally, automatically or imported from a real world
map. Manual generation of the map requires inputs
of two types of information, nodes and edges. A
”node” is one particular point on the map which
can be either a junction or the dead end of the
roads. Furthermore, the junction nodes can be
either normal road junctions or traffic lights.
The edge is the road that connects two points
(nodes) on a map. The attributes associated with
an edge include speed limit, number of lanes, the
Finally, one of the major overhead before road priority and the road length. Figure 5 shows
one can start conducting research using simula- snapshots of nodes editor and edge editor. We
tions is to learn about the internal details of the have also integrated Google Earth into MOVE to
simulator and write customized simulator-specific facilitate the creation of nodes in a realistic set-
scripts to generate various simulation scenarios ting. Google Earth is a tool that enables its user
for the research problem under study (Floyd et to view the satellite image map of any place on
al., 2001). earth. One of the functionality that Google Earth
To reduce such an overhead, MOVE provides provides is called “placemark” which allows the
an interface to automatically generate simulation user to put a mark on any location of the Google
scripts on the fly based on the parameters that the Earth map. Each placemark contains the longitude
user inputs into MOVE. We currently support and latitude information of the selected locations
auto-generation of ns-2 and qualnet simulation and can be saved into a file in KML format (K.
scripts, as shown in Figure 4. tutorial). Hence, one can define the node location
on the Google map and then extract the node
information by processing the saved KML file.
This allows MOVE users to generate a map for
Figure 4. Traffic model generation
358
MOVE
Figure 5. Road map generation
any real-world road on earth for their simulations. inter-departure time of the vehicles originating
Figure 6 shows an example of using Google Earth from the starting road. In addition, a MOVE user
to generate nodes for the major intersections in can define the probability of turning to different
the Eastern Suburb of Sydney.Figure 5. Road directions at each junction (e.g. 0.5 to turn left,
map generation 0.3 to turn right and 0.2 to go straight) in the
The road map can also be generated auto- editor. Figure 8(a) shows a snapshot of the Flow
matically without any user input. Three types of definition Editor.
random maps are currently available: grid, spider, One can also generate vehicle movement
and random networks. There are some parameters manually using the Vehicle Movement Editor
associated with different types of random maps which allows users to specify several properties
such as number of grids and the number of spider of vehicle routes including the number of vehicles
arms and circles. Finally, one can also generate a in a particular route, vehicle departure time, origin
realistic map by importing real world maps from and destination of the vehicle, duration of the trip,
publicly available database. We currently sup-
port the TIGER maps which are available from
U.S. Census Bureau. Figure 7 shows a grid map Figure 6. Generating realistic map using Google
generated from the random map generator and a Earth
street map in the Houston area based on a TIGER
database file.
Vehicular movement editor
The movements of vehicles can be generated

automatically or manually using the Vehicle
Movement Editor. To generate vehicle movement
automatically, one needs to first define a vehicle
flow which describes a fleet of vehicles toward
the same direction. The parameters of each flow
consist of the starting road and destination of the
vehicle fleet, the time to start and end the vehicle
flow, the number of vehicles in the flow and the
359
MOVE
Figure 7. Road map generation using the Map Editor
vehicle speed (including acceleration, deceleration and is also being viewed as a viable commercial
and maximum speed), etc. solution for extending Internet services to public
Figure 8(b) shows a snapshot of the Vehicle transport passengers. MOVE allows users to enter
Movement Editor. Note that, in addition to simu- the bus time table to simulate the movements of
lating vehicle-to-vehicle communication, our tool public transport. We model buses as a group of
is also useful for simulations of vehicleto-infra- vehicles which have similar parameters such as
structure communication (e.g. the communication speeds, routes, etc, associated with it as other ve-
between mobile nodes and road-side static gateway hicles. In addition, to model the bus time table, one
nodes). A static node can be created in MOVE by should define the departure times of the first and
assigning the vehicle with a maximum speed of the last bus and the bus inter-arrival time (which
zero in the Vehicle Movement Editor. is assumed to be constant in our implementation).
On-board communication has recently be- Figure 9 shows the editor for entering the bus route
come an increasingly popular research topic. A information. MOVE goal is to inject the simula-
new paradigm of Networks in Motion is quickly tion with as much detail as possible to provide
attracting interest from the research community a ”realistic” mobility model. We find mobility
Figure 8. Vehicle movement generation
360
MOVE
Figure 9. Bus route generation Figure 10. Comparison between Random Way-
point and model generated by MOVE
packets of a size 64 bytes at the rate of 4 packets

model acts a leading role in VANET, it determine per second. All nodes use 802.11 MAC operating
the nodes location, density, and direction etcetera at 2Mbps. The transmission range is 250m. The
that influence directly VANET performance. Yet propagation model employed in the simulation
a ”fully realistic” simulation is not possible, in is the log normal shadowing model. We used a
mobility model aspect due to human behavior is path loss exponent 2.56 with standard deviation
hardly description (i.e. mood, sex, age, etc.). In 4.0 based on real-world measurement data from
addition, there have a lot of accident in the real an inter-vehicle experiment we previously carried
world, it also changed mobility model. Therefore, out in Sydney suburban area. The road topology
simulation designers must restrict the level of detail generated by MOVE is based on the TIGER
in some place. The challenge is to recognize what database data.
level of detail does not influence answers to the Figure 10 shows the packet delivery ratio of
design questions at hand. AODV with different number of traffic sources.
Each data point represents the average of six runs
and the error bars represent the range of observed
eValuaTion packet delivery ratios. Overall, the packet delivery
ratios increase as the number of traffic sources
In this section, we evaluate the impact of mobility increases, which suggest a higher density of nodes
models generated by MOVE on the performance can increase the network performance as long as
of ad-hoc routing protocol. First, we compare the increasing density does not create more radio
the performance of AODV (Perkins et al., 2003) interference. In addition, the packet delivery ratios
when used with the random waypoint model to of AODV based on MOVE mobility models are
that using the MOVE mobility model. lower than when the Random Waypoint model is
The simulation experiments were carried out used. The results generated from MOVE also have
in ns-2. Each simulation lasts for 900 seconds. larger variations. The larger variance in MOVE
We generated scenarios for 150 nodes moving data points is possibly due to unstable network
in an area of 4 square kilometres. We varied the connectivity imposed by constrained node move-
number of source nodes from 10 to 50, each of ments by roads and traffic control mechanisms
which is a CBR traffic source transmitting UDP (such as traffic lights). Figure 10 clearly shows
361
MOVE
that the simulation results using a more realistic radio transmission range is smaller than the distance
mobility model can be drastically different from between two adjacent clusters. In other words, a
that using a simplistic open field model. Note that link breakage can happen when the inter-cluster
our results are also consistent with prior work distance is larger than the radio coverage.
(Choffnes et al., 2005). Figure 11 shows the distribution of the number
Second, we evaluate the effects of details of of neighboring nodes when ten traffic lights are
mobility models in three case studies. Specifically, included in the simulations. Our results show that
we set out to understand how the existence of traf- each node has twice the number of neighboring
fic lights, driver route choice and car overtaking nodes when traffic lights are simulated, as compared
behavior affect the VANET simulation results. The to the case when traffic lights are not simulated.
number of nodes in our simulations is 300 and the Here we define a “neighboring node” as the node
simulation time lasts for 1000 seconds. The roads which is within the radio range of a vehicle. Hav-
created in the simulation have two lanes. ing a larger number of neighboring nodes typically
suggests a better network connectivity.
Traffic lights simulation As shown in Figure 12, the packet delivery ratio
is improved when the traffic lights are simulated.
In real world, traffic lights are used to regulate traffic Note that in this simulation the distance between
flow moving in different directions. The existence
of traffic lights tends to create a “clustering” effect.
In other words, places where there is a traffic light Figure 12. Effect of traffic light
are likely to have a higher node density due to that
vehicles are forced to stop at the traffic light to wait
for the light to turn green. Intuitively, a high node
density might improve the network connectivity.
On the other hand, a higher node density might
also suggest a higher chance for packet collision
since more nodes might be transmitting at the same
time. In addition, the distance between two adjacent
traffic lights can have a significant effect on the
network connectivity. Specifically, the network
can be “fragmented” by the traffic lights when the
Figure 13. Effect of inter-traffic-light distance

Figure 11. Clustering effect due to the traffic light
362
MOVE
two adjacent traffic lights is shorter than the given overtaking simulation
radio range. In addition, we observe that the number
of packet collisions increase as we increase the In real world, a faster vehicle can overtake some
number of traffic sources. As a result, the packet other slower ones when overtaking is allowed on
delivery ratio decrease when there are more traf- a multi-lane road. Overtaking behavior can have a
fic sources. great effect on the network topology and should be
To understand the effect of inter-cluster dis- considered. Specifically, when overtaking behav-
tance on the simulations results, we increase the ior is not allowed, it usually results in a chain-like
distance between two adjacent traffic lights (from topology and a shorter and uniform inter-vehicle
200m to 400m) so that the inter-cluster distance is distance (the uniform distance is due to that the
larger than the effective radio distance. As shown vehicle needs to maintain a safe distance from
in Figure 13, in this scenario we observe frequent the adjacent cars), which often suggests a better
link breakage between two adjacent clusters which network connectivity. We observe a dramatic
significant degrade the network performance. impact on the network performance when the
The effective radio range is around 250m in this overtaking behavior is allowed. In addition, we
experiment. find that the effect of overtaking behavior is less
significant when the network density is higher.
Turning simulation As shown in Figure 15, the packet delivery ratios
in overtaking-allowed scenario is close to results
In real world, a driver normally has to decide his of no-overtaking scenario when we increase the
moving direction at an intersection. He can choose number of nodes from 250 to 350.
to either go straight, turn left, or turn right. MOVE In summary, we show that details of mobil-
allows a user to define the turning probability to ity models such as the existence of traffic lights,
different directions at each intersection (e.g. 0.5 driver route choice and car overtaking behavior can
to turn left,0.3 to go straight and 0.2 to turn left) have a drastic impact on the VANET simulation
in the Vehicle Movement Editor. As shown in results. We argue that the faithfulness of simula-
Figure 14, we find that different choices of route tion results is proportional to the realism of the
directions can significantly change the simulation parameters and the models used in the simulations.
results (the x-y-z notation in Figure 14 means that Therefore, selecting appropriate level of details
the car has x% of chance to turn left, y% to go in the mobility model for a VANET simulation is
straight and z% to turn right). a very important yet challenging task.
Figure 14. Effect of driver route choice Figure 15. Effect of car overtaking behavior
363
MOVE
relaTed work follows: random walk (Zonoozi et al., 1997) (this

model not discuss in the paper), RWP, RPGM,
In this section, we have discussed mobile mobility and their approach, mobility vector model. The
model with level of different details and VAENT mobility vector model’s advantage can use for
simulators. Mobility models play a critical role in different behavior, such as gravity model, location
VANET. This is because mobility affects routing dependent model, targeting model, etc. Due to it
and network performance etc. Therefore, the user have deceleration and change direction behaviors
should carefully choose suitable mobility model for more realistic mobility model. This meaning the
in applications. model possesses basic characteristic for vehicular
mobility. But this model still hard widely use in
mobility model with level the real world. Although human drivers usually
of different details is intuitive, it yet need consider different traffic
condition, such as traffic congestion or sparse. For
Random Way-Point (RWP) (David et al., 1996) is instance, when the traffic condition is congestion,
the earlier mobility models used in MANET simu- the vehicles only suit car following. Vice versa,
lation. It consider without obstacles topography overtaking most happen at traffic sparse. There-
and nodes to be able to move freely in simulation fore, this is important phenomena.
area. RWP is widely to use mobility model in Camp et al. (2002) survey variety of the mo-
simulation (Das et al., 2000 ; Holland et al., 1999 bility model in ad hoc networks, such as random
; Perkins et al., 1999,; Johansson et al., 1999.), walk mobility, RWP, random direction mobility,
which only suit mobile nodes in MANET. This is probabilistic random walk etc. This can sort two
because VANET environment have obstacles and main models: entity models and group models
each vehicle restrict by street. Authors of (Hong for mobile node. Fortunately, VANET constrain
et al., 1999) present a Reference Point Group by street that the mobility model become simple.
Mobility (RPGM) model to depict the relationship But in the real world, the driver has indefinite
among mobile hosts. This model based on Random behavior with sex, age, mood, habits etc. effects
Waypoint model to use checkpoint for movement. lead to different mobility pattern (e.g. overtaking,
Although this model without any obstacle, but a car following, etc.). The traffic light also affect
suitable checkpoint scenario can model realistic vehicles movement in VANET. Bai et al. (2003 ;
mobility patterns. Unfortunate, a common and 2003, November) introduced RWP, group mobil-
realistic mobility model found hard. (Bettstetter et ity, Freeway, and Manhattan mobility models. Dif-
al., 2001) present an improved Random Direction ferent mobility models have especially application
Model instead of Random Waypoint Model. It in- and characteristics as follows: group mobility use
troduces a stop-turn-and-go behavior for vehicles in military battlefield or motorcade of travel etc.
or bicycles movement patterns that discuss alike Generally, this mobility model have cluster and
vehicular nodes stop over at intersection. This both speed and direction. Freeway mobility model,
model can extend to turning and stopping traffic this mobility model should have car-following and
light events in VANET. Similar to the previous overtaking movement behaviors in general case,
literature, authors of (Royer et al., 2001) use simu- such that usually happen at traffic congestion or
lation to present random direction model. Due to rarefaction. Nevertheless, the authors not mention
the simulations focus on mobile ad hoc network car following and overtaking. The urban environ-
and without considering realistic map. In the other ment use Manhattan mobility model, the mobility
words, it not suit for VANET. In addition, Hong model is a simple urban mobility model that the
et al. (2001) discuss different mobility model as vehicles only move along the grid of horizontal
364
MOVE
and vertical streets on the map. Actually, the urban show obstacle in VANET network. This result same
have a lot of traffic light to manage and control as previous literature (Sven Jaap et al., 2005) not
traffic flow, this issue discuss in evaluation section. consider traffic light and overtaking.
Until now, we survey the mobility models that Street RAndom Waypoint (STRAW) (Choffnes
most use RWP and without considering obstacle. et al., 2005), it is a random waypoint constraint by
They are not suitable use on VANET. road. The scheme has considered traffic light control
Saha et al. (2004) proposed based-on TIGER and car following model. It is also use shortest path
maps’ macro mobility model. It is only consider- algorithm to calculate movement path. The authors
ing Dijkstra’s single source shortest path algorithm though have considered previous condition. But
from source to destination location. Although it still lack overtaking criterion that cause convey
use dynamic source routing protocol to evaluate effect in street. However, it is not realistic and then
performance by NS2. However, the scheme lacks only uses random method. Mahajan et al. (2006)
traffic light mechanism and overtaking behavior. proposed Stop Sign Model (SSM), Probabilistic
Jardosh et al. (2005) present an obstacle mobility Traffic Sign Model (PTSM), and Traffic Light
model that the placement of obstacles that restrict Model (TLM). This paper focus in traffic control
movement and signal propagation, although it system. In SSM, each vehicle around the intersec-
focuses discussion in mobility model of people, tion must stop at the stop sign in duration time.
but the builds can instead of extracting data from PTSM use a probability p to deice the vehicle stop,
TIGER files. This behavior spreads street restric- this paper use with an empty street queue, but this
tion and build obstacles in VANET environment. model is not coordinating among different direc-
It also use distance dynamic to adjust propagation tions. In other words, PTSM can utilize different
models. This is for realistic simulation and good traffic condition to adjust traffic light time and
idea. (Stepanov et al., 2005) Using probabilities decrease traffic congestion. The TSM similar to
decide trips and use estimated travel time to PTSM, it only changed at coordinating among dif-
choose movement path. This can increase realistic ferent directions. Anyhow, any types’traffic control
of mobility model that the probabilities can draw system both has pause time to control vehicle. This
trips’ peak and off-peak in duration. In addition, is important factor in traffic control system. When
the scheme hard get accurate travel time. This is increasing the pause time at the intersections, each
because VANET fast changed. The travel time easy vehicle waiting time will elongate that increase
affect by traffic congestion or accident. In (Sven traffic congestion. In the other street, this street’s
Jaap et al., 2005), the authors present city mobility vehicles not need stop over at intersection and move
model that build on IDM (Intelligent-Driver Model) forward, this street almost no waiting time. This is a
(Treiber et al., 2000) with probability turning. It trade-off in different application. The authors’simu-
is meaning the model that support car following lation results point out long pause-time which can
and turning. But it cannot provide network simu- increase packer delivery ratio. This only consider
lator traces, without considering traffic light and at street of local. This case is unrealistic. In (Potnis
overtaking. Therefore, this model not enough call et al., 2006), the authors have simulated SSM and
realistic mobility model. Zimmermann et al. (2005) TSM mobility model to compare packet delivery
proposed a voronoi-based mobility model for urban ratio. On the other hand, (Marfia et al., 2007) Us-
environments, this model is using a spatial area ing CORSIM (CORSIM, 2009) and TRANSIMS
obtained movement paths which are computed by (TRANSIMS) generate different mobility model
the multiple application of Voronoi graphs. Due to in downtown, and then use Qualnet to compare
the vehicles movement is constrained to street by network performance with vary transmission range
the voronoi graph, this phenomenon fit VANET etc. In addition, the authors also have discussed with
network. In principle, this mobility model only infrastructure (i.e. AP) to improve packet delivery
365
MOVE
ratio. Nevertheless, the context lack more detail of sic motion constraints and does not consider any
mobility model discussion, such as overtaking and micro-mobility. Furthermore, BonnMotion is a
lane change impact. text-based application that runs on a command shell
and does not provide any graphical user interfaces
VaneT simulators as MOVE does. Complementary to these previous
efforts, our work emphasizes on creating a tool that
Groovesim (Mangharam et al., 2005) is a topog- allows users to rapidly generate realistic mobility
raphy-accurate street-map based vehicle network models for VANET simulations.
simulator and is based on GrooveNet, a geographic
routing protocol for vehicular networks. It provides
several different modes of operation. In Drive conclusion and FuTure work
Mode, GrooveSim can process data from a GPS
unit to provide a real-time map of the vehicle’s In this chapter, we describe a tool MOVE which
current location. It can also be used as an emulator is based on an open source micro-traffic simulator
in Hybrid Simulation Mode where real vehicles on SUMO. MOVE allows user to quickly generate
the road and virtual vehicles in the simulation can realistic mobility models for vehicular network
interact with each other. Groovesim also provides simulations. In addition, MOVE provides an inter-
a tool for analyzing the simulation results. One face to automatically generate simulation scripts
limitation of Groovesim is that it is strongly tied for ns-2 and qualnet. Finally, we show that the
to one specific routing protocol (i.e. GrooveNet), simulation results using MOVE is significantly
which limits its use for simulating other routing different from that using the commonly used ran-
protocols in a VANET environment. In addition, dom waypoint model. Moreover, we have offered
GrooveSim does not provide mobility traces for traffic light, turning, and overtaking cases studied in
network simulators. VANET simulation. Each case have different effect
STRAW is an extension of SWANS (Scalable in simulation, thus selecting a suit level of detail
Wireless Ad Hoc Network Simulator) (Haas et for a simulation is key point. MOVE is publicly
al., 2005), a Java-based simulator for wireless available and can be downloaded via the following
simulations. STRAW contains simulation tools for URL - http://lens1.csie.ncku.edu.tw/MOVE/.
generating mobility models and traffic models and
is also able to use real street maps like TIGER data
to build the road topology. However, currently the reFerences
mobility models can be supported by STRAW is
limited. For example, while STRAW supports mul- B. A. mobility scenario generation and analysis tool,
tiple lanes, the vehicles are not allowed to change http://web.informatik.uni-bonn.de/iv/mitarbeiter/
lane and the starting position is not configurable. dewaal/bonnmotion/.
Another drawback of this tool is its dependency on
Bai, F., Sadagopan, N., & Helmy, A. (2003) “Im-
SWANS. Finally, STRAW does not provide any
portant: a framework to systematically analyze
GUI that allows the users to visualize the move-
the impact of mobility on performance of routing
ments of cars.
protocols for ad hoc networks,” in The 22th IEEE
BonnMotion (B. A. mobility scenario genera-
Annual Joint Conference on Computer Commu-
tion and analysis tool) is a simple tool that can be
nications and Networking INFOCOM 2003 (pp.
used to create and analyses mobility scenarios.
825–835).
Similar to MOVE, the mobility scenarios cre-
ated by BonnMotion can be exported to ns-2 and
qualnet. However, BonnMotion only models ba-
366
MOVE
Bai, F., Sadagopan, N., & Helmy, A. (2003, No- Haas, Z. J., & Barr, R. (2005) “Density-independent,
vember) “The important framework for analyzing scalable search in ad hoc networks,” in Proceedings
the impact of mobility on performance of routing of IEEE International Symposium on Personal In-
for ad hoc networks,” AdHoc Networks Journal - door and Mobile Radio Communications. [Online].
Elsevier Science, vol. 1, no. 4 (pp. 383–403). Available: http://jist.ece.cornell.edu/docs/050911-
density.pdf
Bettstetter, C. (2001) “Smooth is better than sharp:
a random mobility model for simulation of wireless Heidemann, J., Bulusu, N., Elson, J., Intanagonwi-
networks,” in Proc. MSWiM01, ACM. wat, C., Lan, K. C., Xu, Y., et al. (2001) “Effects of
detail in wireless network simulation,” in Proc. of
Camp, T., Boleng, J., & Davies, V. (2002). “A survey
Communication Networks and Distributed Systems
of mobility models for ad hoc network research,”
Modeling and Simulation Conference, Jan. 2001.
in Wireless Communications and Mobile Comput-
[Online]. Available: http://www.isi.edu/johnh/
ing (WCMC): Special issue on Mobile Ad Hoc
PAPERS/Heidemann00d.html
Networking: Research . Trends and Applications,
2(5), 483–502. Holland, G., & Vaidya, N. H. (1999) “Proc. mobi-
com, acm intern. conf. on mobile computing and
Choffnes, D. R., & Bustamante, F. E. (2005) “An
networking,” in Proc. IEEE Infocom.
integrated mobility and traffic model for vehicular
wireless networks,” in Proc. of the 2nd ACM Inter- Hong, X., Gerla, M., Pei, G., & Chiang, C.-C.
national Workshop on Vehicular Ad Hoc Networks (1999) “A group mobility model for ad hoc wire-
(VANET), Cologne, Germany. [Online]. Available: less networks,” in Proc. ACM MSWiM.
http://www.aqualab.cs.northwestern.edu/publica-
Hong, X., Kwon, T., Gerla, M., Gu, D., & Pei, G.
tions/DChoffnes05vanet.html
(2001, January) “A mobility framework for ad
CORSIM. (2009). http://www.fhwa-tsis.com/. hoc wireless networks,” in Proc. of ACM Second
International Conference on Mobile Data Manage-
CORSIM. http://mctrans.ce.ufl.edu/featured/tsis/
ment (MDM 2001).
version5/corsim.htm.
Jardosh, A. P., Belding-Royer, E. M., Almeroth,
Das, S. R., Perkins, C. E., & Royer, E. M. (2000)
K. C., & Suri, S. (2005) “Real-world environment
“Performance comparison of two on-demand
models for mobile network evaluation,” in In the
routing protocols for ad hoc networks,” in Proc.
IEEE Journal on Special Areas in Communications
IEEE Infocom.
- Special Issue on Wireless Ad hoc Networks.
David, B., & David, A. (1996) “Dynamic source
Johansson, P., Larsson, T., Hedman, N., Mielczarek,
routing in ad hoc wireless networks,” in Mobile
B., & Degermark, M. (1999) “Routing protocols
Computing, vol. 35, Kluwer Academic (pp.
for mobile ad-hoc networks - a comparative
153–181).
performance analysis,” in IEEE MOBICOM, p.
Floyd, S., & Paxson, V. (2001) “Difficulties in simu- 195V206. K. tutorial, http://www.keyhole.com/
lating the Internet,” ACM/IEEE Transactions on kml/kml tut.html.
Networking, vol. 9, no. 4 (pp. 392–403). [Online].
Mahajan, A., Potnis, N., Gopalan, K., & Wang, A.
Available: http://www.icir.org/vern/papers.html
A. (2006) “Urban mobility models for vanets,” in
in Proc. of the IEEE Workshop on Next Generation
Wireless Networks (WoNGeN).
367
MOVE
Mangharam, R., Weller, D. S., Stancil, D. D., Ra- SimulationP. M. T. (2009). http://www.paramics-
jkumar, R., & Parikh, J. S. (2005) “GrooveSim: online.com/.
A topography-accurate simulator for geographic
SimulatorO. (2009). http://www.opnet.com/.
routing in vehicular network,” in Proc. of the 2nd
ACM International Workshop on Vehicular Ad Hoc SimulatorQ. N. (2009). http://www.scalable-
Networks(VANET),Cologne,Germany.[Online]. networks.com/.
Available: http://www.andrew.cmu.edu/user/ra-
hulm/Research/Pubs/ vanet05 mangharam.pdf Stepanov, I., Marron, P.-J., & Rothermel, K.
(2005l) “Mobility modeling of outdoor scenarios
Marfia, G., Pau, G., Giordano, E., Sena, E. D., for manets,” in Proceedings of the 38th annual
& Gerla, M. (2007) “Evaluating vehicle network Symposium on Simulation ANSS ’05, Apr. 2005
strategies for downtown portland: Opportunistic (pp. 312–322).
infrastructure and importance of realistic mobility
models,” in proc. of MoBiOpp 2007, co-located Sven Jaap, M. B., & Wolf, L. (2005) “Evaluation of
with ACM/USENIX International Conference routing protocols for vehicular ad hoc networks in
on Mobile Systems, Applications, and Services city traffic scenarios,” in in Proc. of the 5th Inter-
(MobiSys 2007). national Conference on Intelligent Transportation
Systems Telecommunications (ITST).
P. simulation VISSIM. (2009). http://www.english.
ptv.de/. T. N. S. ns 2.(2009). http://www.isi.edu/nsnam/
ns/index.html.
Perkins, C., Belding-Royer, E., & Das, S. (2003)
“Ad hoc on demand distance vector (aodv) rout- TIGER (Topologically Integrated GEographic
ing,” rFC 3561. Encoding and Referencing). (2009). http://www.
census.gov/geo/www/tiger/.
Perkins, C. E., & Royer, E. M. (1999) “Ad-hoc
on-demand distance vector routing,” in Proc. TRANSIMS. http://transims.tsasa.lanl.gov/.
IEEE Workshop on Mobile Computing Systems Treiber, M., Hennecke, A., & Helbing, D. (2000)
and Applications. “Congested traffic states in empirical observations
Potnis, N., & Mahajan, A. (2006) “Mobility models and microscopic simulations,” IPhys. Rev., vol.
for vehicular ad hoc network simulations,” in proc. 62, no. 2.
of the 44th annual Southeast regional conference, Zimmermann, H.-M., Gruber, I., & Roman, C.
ACM (pp. 746–747). (2005) “A voronoi-based mobility model for urban
Royer, E. M., Melliar-Smithy, P. M., & Mosery, environments,” in in Proc. Of the European Wire-
L. E. (2001) “An analysis of the optimum node less 2005 (EW05) .
density for ad hoc mobile networks,” in Proc. of Zonoozi, M. M., & Dassanayake, P. (1997).
the IEEE International Conference on Communi- User mobility modeling and characterization of
cations, Jun. 2001. mobility patterns . IEEE Journal on Selected
S. S. of Urban Mobility. (2009). http://sumo. Areas in Communications, 15(7), 1239–1252.
sourceforge.net/. doi:10.1109/49.622908
Saha, A. K., & Johnson, D. B. (2004) “Modeling

mobility for vehicular ad hoc networks,” in Proc. of
the 2nd ACM International Workshop on Vehicular
Ad Hoc Networks (VANET). [Online]. Avail-
able: http://www.cs.rice.edu/amsaha/ Resume/
vanet2004/vanet2004.pdf
368
369
Chapter 22
Security Attacks of
Vehicular Networks
Jen-Chun Chang
National Taipei University, Taiwan, R.O.C.
Chun-I Fan
National Sun Yat-sen University, Taiwan, R.O.C.
Ruei-Hau Hsu
National Sun Yat-sen University, Taiwan, R.O.C.
absTracT
The application of vehicular ad hoc network (VANET) improves driving safety and traffic management.
Due to the above applications, security attacks on VANET can be serious threats all the time. VANET
is a special form of mobile ad hoc network (MANET). Hence any attacks exist on MANET also can be
arisen on VANET. Moreover, some special attacks can be raised on VANET, which do not exist on MANET.
Nevertheless, some characteristics of VANET can be positive effects and some can be negative effects
on security issues. Before designing the security mechanism to defend attacks, the authors should take
the positive effects and avoid the negative effects on the security of VANET. Furthermore, the authors
class all possible attacks of VANET from every network layer. They also introduce the reason of forming
every attack and the possible effect on VANET in detail. Therefore this chapter helps understanding the
latent threats and the useful resources of security issues on VANET.
inTroducTion million people injured, and the costs associated with

traffic accidents estimated 160 billion euros. The
In recent years, people have fixed their eyes upon annual costs associated with crashes (like hospital
traffic-safety topic, because current traffic accident bills and damaged property) total nearly 3 percent
statistics are notoriously horrific. According to the of the world’s gross domestic product (GDP) in
European Red Cross Road Safety Campaign report, 2000, or roughly US $1 trillion. In order to reduce
approximately 43,000 people die every year on the the traffic accidents, governments and researchers
roads of the European Union (EU), with around 1.8 around the world try to find out effective solutions,
and Vehicular Communication (VC) system, or we
can also call it Vehicular ac hoc network (VANET),
DOI: 10.4018/978-1-60566-840-6.ch022
Security Attacks of Vehicular Networks
seems to be one of the answers and will be ubiq- • Limited physical access: The operator of
uitous everywhere in the not-too-distant future. IVC usually limited to the owner of car or
Vehicular ac hoc network (VANET) is likely authorized personnel.
to become the most universal and relevant form • Periodic maintenance: Cars always
of ad hoc networks due to the urgent need of need to be maintained in a period of time.
driving safety. Another reason about the fast Therefore, the IVC also can check and up-
development of VANET is the impact to the date regularly.
market. There are more than 50 applications have • Secure computing platform: Automotive
been submitted by major car manufactures like environment it seems inevitable that some
BMW, Daimler-Chrysler, Ford, and GM which kind of secure computing platform must be
available in the future.
are based on Dedicated Short Range Communica-
tion (DSRC) technology. DSRC is a short range
Properties that have a negative effect:
wireless protocol specifically for automotive use.
It offers communication between vehicles and
• High mobility: High degree of mobility is
Road Side Units (RSUs). This technology for
one of properties of vehicles. It means that
VANET applications is working in the 5.9 GHz
the average speed of the node of VANET
band (U.S) or 5.8 GHz band (Japan, Europe).
will be very high and the average connection
VC system includes two types of commu- time will be very short. Therefore, when we
nications: design the security mechanism, the commu-
Inter-vehicle communication (IVC) (or some- nication time and computation time should
one call it Vehicle-to-vehicle (V2V) communica- be considered.
tion) and Roadside-to-vehicle communication • Large number of nodes: IVC network can
(RVC) (or someone call it Vehicle-to-infrastruc- be a huge ad hoc network. Scalable solu-
ture (V2I) communication). All two types are tions for adequate and sufficient perfor-
based on wireless multi-hop communication. mance should be considered.
• No centralized infrastructure: When we
inter-Vehicle communication deal with a distributed ad hoc network, the
centralized infrastructure is only available
IVC systems have some properties that support at specific situations. The design of some
security and others that are negative effect. security building block should be adapted
Properties that have positive effects. to such kind of infrastructure, such as trust
management and key distribution and re-
• No energy constraints: unlike the sen- quires new concepts.
sor node in ad hoc network and/or sensor • Privacy concerns: Privacy is a serious
network and some mobile device, such as problem in IVC system because cars are
cellular phone and personal digital assis- highly personal devices and the owners
tant (PDA), cars usually provide enough will keep it for a long duration. The system
energy to operate communication system design should reflect the need for flexible
and related computation of security. identifiers.
• No user interaction: The scenario of IVC
• Known position and time: This informa-
system is that no user interaction possible
tion is required for most safety applica-
since it could distract drivers and reduce the
tion on VANET. This information can also
popularity and usability of IVC system.
be used to security application.
370
One of famous IVC system model is pro- IVC as communication only, all connections must
posed by the network on wheels (NoW) project. be established by multi-hop communication, it
The NoW system is a generic model for IVC is a huge works and consume large amount of
architecture. This model consists of four major resources. That is the reason we must use IVC
aspects. First, the radio channel and the protocol and RVC interchange. We can also connect to
are used on NoW system. Second, the hardware Internet through RVC and get other multimedia
and the software running on NoW. The three major services, like on-line TV, on-line game …etc. That
kinds of platform are within the NoW system, make everything come true and vehicles are not
one is onboard units (OBU), that are installed just vehicles, they become to be entertainment
in the vehicles, road side units (RSUs) as parts centers.
of some road infrastructure, and the HomePC or We will introduce VANET characteristics,
commercial Access Points providing content or architecture, applications and its security attack
Internet access. Third, the input of sensor to the separately. VANET is a special or even excep-
different processing units, that can be all kinds of tional case of mobile ad hoc network (MANET).
physical sensors, such as temperature, oil on the MANET is one kind of wireless ad-hoc network,
road and so on. Finally, the security infrastructure it is not only include the same characteristics as
behind the NoW system consists of some ele- wireless ad-hoc network, such as an open peer-to-
ments, like the vehicle manufacturers, certifica- peer network architecture, lack a central instance,
tion authorities, traffic authorities, and certified each node is willing to forward data for other
staff. The following figure (Figure 1) is a generic nodes, self-configuration and self-maintenance
system model of NoW. capabilities. Fully distributed nodes may move
Roadside-to-vehicle communication (RVC) arbitrarily and change their topology frequently
is used to complement IVC. If VANET just use without prior notice.
Figure 1. The architecture of NoW system
371
The archiTecTure oF The infrastructure domain provides connectivity

Vehicular neTwork to Internet nodes. The vehicles can get applications
like multimedia services through RSUs to Internet. If
VANET architecture is composed of two kinds the vehicle we want to communicate is out of range
of communicating nodes, including vehicles that multi-hop can reach, we can use OBU to connect
(nodes) and Road Side Units (RSUs). Vehicles to RSU of ad hoc domain we located, and the RSU
can also be split into two parts: private (belonging will communicate to RSU where the vehicle we want
to individuals or private companies) and public to connect located, and then we can access to the
(public transportation vehicles, e.g., buses, and vehicle as two RSUs as intermediate nodes.
public services)
The network architecture can roughly split into
three domains (shown in Figure 2): the in-vehicle The applicaTions oF
domain, the ad hoc domain and the infrastructure Vehicular neTwork
domain. The in-vehicle domain is a sub-network
(of the VANET) with an On Board Unit (OBU) and VANET could provide several safety applications
several Application Units (AUs), such as Personal (Lin et al., 2007) to protect driver and passenger
Digital Assistant (PDA), mobile phone… etc. and ubiquitous multimedia services. These safety
The ad hoc domain is composed of multiple applications include Electronic Emergency Barke
OBUs and RSUs along the road. OBUs and RSUs Light (EEBL), Road Hazard Condition Notifica-
are equipped with IEEE 802.11-like wireless tech- tion (RHCN), Road Feature Notification (RFN),
nology. They can either communicate each other Slow Vehicle Alert (SVA), and Post Crash Noti-
directly or through multi-hop communication. fication (PCN). These safety applications could
Figure 2. VANET architecture
372
Figure 3. High-level system architecture view
help driver the dangerous situation or the condi- MANET faced and owned. Watch out – there still
tion of traffic to void possibly dangers or threats. exists some VANET-specific attacks.
EEBL notifies drivers when a vehicle is closed to There is one thing we MUST keep in mind
them and rapidly decelerates. It could avoid the that security attack is cross-layer issue rather
chance of accident of rear-end collisions. RHNC single layer issue. We can totally understand the
broadcast the information of road hazard, such as truth from Table 1 below. The security attacks for
snow accumulation or trash. RFN alerts drivers VANETs span the entire network protocol and not
the landform or sections with special limitation just application layer
they will pass. If drivers approach a steep hill, VANET security attacks includes
school, they may need to slow their speed with a
must lower ones to avoid accidents. SVA and PCN 1. Viruses, worms and malicious code
alert drivers a slow vehicle or a possible crash in 2. Bogus information attack
the lanes ahead. The information of such safety 3. Message modification attack
applications are come from all nodes of VANET. 4. ID disclosure of other vehicles
Due to these safety applications, drivers have more 5. Movement tracking
time to reach to avoid some dangers or accidents 6. Denial of Services (DoS)
that may happen with huge damages. 7. Impersonation attack
8. Cheating with positioning information
9. RSU replication attack:
The aTTacks oF 10. Unauthorized preemption attack
Vehicular neTwork 11. WEP vulnerability
12. Routing attack
VANET is a special or even exceptional case 13. Data packet forwarding attack
of MANET, it includes all security attacks that
373
Table 1. The security attacks for VANETs span the entire network protocol stack
Layer Security Attacks

Application Layer Viruses, Worms, Malicious Codes, Bogus information attack, message modification attack, Movement tracking, DoS…etc.
Network Layer Ad hoc routing attack, data packet forwarding attack, ID disclosure attack, Movement tracking, DoS…etc.
Data Link Layer WEP vulnerability, ID disclosure attack, Movement tracking, DoS…etc
Physical Layer DoS attacks
We will explain each attack behavior care- a loophole, and malicious user may produce fake
fully below. or untrue messages to achieve certain goals. In
Viruses, worms and malicious code: Viruses, VANET circumstance, bogus information may
worms and malicious code are tools used to invade even cause more critical consequences, like traf-
OBUs or RSUs. After intrusion is successful, fic jam or traffic accidents. For example: one
malicious user can get personal information of may send a fake traffic jam message to the other
OBU owner, and convey some fake information drivers and affect their behavior in order to divert
to disturb the traffic. Therefore, the manager of traffic from a given road and get a better traffic
VANET system can have some strategy to avoid the condition (See Figure 4).
damages caused from virus, worms and malicious Message modification attack: Some people
code, such as the choice of operating system, the might consider that message modification attack
periodic update and checking, security policy to is similar to bogus information attack, but they
access operating system and so on. It is also helpful are totally different. Bogus information attackers
to prevent such kind of threat to break VANET produce messages which are purely fictitious, but
system by policy or standard procedure. message modification attackers just alter rather
Bogus information attack: In MANET envi- create messages during or after transmission. The
ronment, all messages are broadcast even though message modification attacker may wish to change
messages which have definitely destination for the source or content of the messages.
private use. Data is transmitted via air interface ID disclosure of other vehicles: In VANET
through shared wireless medium. It is obviously architecture, data is transmitted via air interface.
Figure 4. The in-vehicle domain
374
And multi-hop transmission strategy makes more less multi-hop communication or the Internet via
and more personal information is available in the the TCP/IP protocol stack. Every vehicle or RSU
infrastructure. According to the above two reasons, will be allocated a unique and static IP and MAC
the probability of personal information disclosure addresses by authority. The addresses represent
go up quite substantially. The attackers can get node’s identity, and researchers try to hide the
users’ personal information through eavesdrop- information preventing attackers capture it. If
ping or accessing the infrastructure illegally. We node’s IP and MAC addresses have been disclosed,
call this attack technique as ID disclosure of other user’s privacy no longer exists and user’s move-
vehicles, also known as the Big Brother scenario. ment will be tracked by attackers.
We must keep one thing firmly in mind that ID In Figure 5, we show header of routing packets,
disclosure attack is cross-layer security issue. In it includes immutable and mutable field. Immu-
order to authenticate client’s identity, every ap- table are those fields that unchanged from sender
plication service will execute an authentication to destination even through intermediate hops
protocol to client. If user lose his identity, make forward the packet to the destination. Mutable
his identity public or eavesdropped by malicious fields are allocated to be altered by intermediate
users, attackers can masquerade by using this nodes. We can see operation on these fields in
identity, and make victims to pay high payment Figure 6.
bill or let them as scapegoat. The attack behavior We can induce that immutable fields are used
belongs to application layer of network protocol to store IP addresses, and mutable fields are used
stack. to store MAC addresses. The whole routing packet
As we mentioned above, vehicular communi- is not protected by encryption. Although malicious
cation is based on IEEE 802.11-like technology. user cannot disclose identity of user from applica-
The applications of VANET can either use wire- tion layer message, he can still catch information
Figure 5. Example of bogus information attack, attacker disseminate false information to affect the
decisions of other vehicles and thus clear the way of attacker
375
Figure 6. Example packet type with mutable and immutable fields (mutable fields are red, and immutable
fields are blue)
from routing packet. We call this attack as IP and we must keep in mind that DoS is a cross-layer
MAC disclosure attack. issue, from message replay attack, layer 2 packet
flooding, Radio-Frequency (RF) interference to
movement Tracking jamming.
A message replay attack means that a valid data
Movement tracking goal can be done by two transmission is repeated or delayed maliciously.
ways. An adversary intercepts the data from the origi-
nator and retransmits it. It may cause DoS and
1. Malicious user gets victim’s identity through paralysis network or the attacker can impersonate
ID disclosure attack first, and then he can other user successfully.
proceed to track victim’s movement. Every nodes of VANET want to determine
2. Malicious user is in the area where victim’s every packet’s source and destination must to
radio range covered. As victim move, at- dispatch packet and check MAC address or IP
tacker also moves with the victim. If attacker address. The dispatch action must take computa-
never moves out of the radio range of victim, tion resource. If a node receives huge amount of
he can track victim’s movement forever. This packets, then he will be down due to resources
way of movement tracking is like ‘physical’ are exhausted.
way. RF (or Electromagnetic interference, EMI)
is a disturbance that affects an electrical circuit
Denial of service (DoS): Denial of service at- due to electromagnetic radiation emitted from an
tack or distributed denial of service attack (DDoS external source. The disturbance may interrupt,
attack) is an attempt to make a computer resource obstruct, or otherwise degrade or limit the effec-
is unavailable to its clients. We can simply explain tive performance of the circuit.
it as an adversary broadcasts irrelevant bulk mes- Impersonation attack: Impersonation attack
sages to the VANET where he located in, taking is also known as Masquerade. As implied by the
up the cannel and consume the computational name, adversary actively pretends to be another
resource of the other nodes. There is one thing vehicle or even RSU by using false identities to
376
fool the others. The attacker may try to achieve WEP: Wired Equivalent Privacy (WEP) is
malicious or rational objectives. A malicious at- an algorithm to secure IEEE 802.11 wireless net-
tacker seeks no personal benefits from the attacks works. Because wireless networks communication
and aims to harm the members or mal-function is based on broadcast through radio, it is easier to
the network for no reason. That is why a mali- eavesdrop than wired networks. That is why we
cious attacker is unpredictable, because his attack need encryptosystem to provide confidentiality as
is disregarding costs and profits. All he wants is wired networks, and WEP is named as this reason.
just contend to appetite for destruction. On the But since 2001, cryptanalysts find several serious
contrary, a rational attacker seeks personal profit weaknesses which will let WEP to be cracked by
and hence is more predictable in terms of the at- available software about few minutes. Now we
tack means and the attack target. can assure that WEP is vulnerable.
Cheating with positioning information: Routing attacks: The family of routing attack
Cheating with positioning information is a special is defined as any action of notifying “wrong” rout-
case of message modification attack. The attackers ing updates (which means not follow the speci-
just alter the positioning information field of the fications of the routing protocol). For example,
message like perceived position, speed, direction, spoofing, forging of routing signaling messages
and so on. Cheating with positioning information …etc. are routing attacks. These behaviors may
may lead to routing attack, because position-based cause some specific attacks which are related to
routing protocol is the most popular routing al- the VANET, e.g: We will introduce some specific
gorithm in VANET environment. If routing table attack behaviors which are related to the VANET,
is not correct, data packet cannot be forwarded e.g.: routing loop and malfunctioning of the net-
correctly, and messages may lost (e.g. sinkhole) work, sinkhole attack…etc. We will describe in
or cause routing loops. Data packet forwarding attack.
RSU replication attack: Surveying the de- Data packet forwarding attack: The at-
ployments of VANET, we can find out that there tackers cause the data packets to be delivered in
exist a large number of RSUs to complement a way that is intentionally inconsistent with the
shortcomings of multi-hop communications. If routing states. Another type of packet forward-
protection of RSUs is not sufficient, they might ing attack is the Denial-of-service (DoS) attack
be compromised by malicious users. Afterwards, via network-layer packet blasting, in which the
attackers can use these captured RSUs to launch attacker injects a large amount of junk packets
any malicious attack, such as bogus information into the network.
attack and unauthorized preemption attack (as Because ad hoc routing protocols are various,
we described below), etc. Maybe we can say we will focus on position-based routing which
RSU replication attack is basis of unauthorized is one of most popular used to construct routing
preemption attack. table. The position–based routing protocol policy
Unauthorized preemption attack: The is – a node forwards a given packet to one one-
adversary may take control a RSU, especially a hop neighbor that is closest to the destination than
traffic light, and then make it to provide special others. Although position based routing protocol
traffic priority for specific vehicles. It is similar (one kind of geographical routing protocols) is
to RSU replication attack common used, it is still inability to against spoof-
After introducing upper layer and cross-layer ing and forging of routing signaling messages
attacks, we will discuss lower layer (such as net- like other ad hoc routing protocols and results
work and link layer) attacks below: in creation of routing loops and mal-function of
the network.
377
Figure 7. Operations on mutable and immutable fields
Figure 8. Two exemplary attacks in position-based routing protocol
In Figure 7: The left figure shows a regular conclusion

forwarding process from the source via forward-
ers (vehicles) V1, V2 and V3 to the destination. In this chapter, we have shown security attacks
In the middle figure, the attacker Att forges its in VANET environment. We first briefly describe
location as Att’. It makes V2 selects the attacker VANET characteristics, architectures, applica-
as next hop as next hop in the forwarding chain tions. After constructing the common sense, we
because its position is closer to destination. The elaborate all possible security attacks in VANET.
attacker receives messages (that should be sent Because VANET is a special case of MANET, it
to next hop) and drops all messages. It makes all includes all MANET security attacks and some
packets lost, and we call it ‘sinkhole’ attack. In special attacks only happened in VANET environ-
the right figure, the attacker forges its position as ment. Maybe someone will ask why we need to
Att’ and V2 selects the attacker as next hop in the know attacks in VANET, because VANET seems
forwarding chain. The attacker receives all mes- to become the most relevant network around the
sages and forwards the received messages back world and VANET includes an important com-
to the previous node V2. It results routing loop. ponent – human beings. The basis of VANET
Messages will be sent back and forth but can’t be development is to protect drivers from accidents,
sent to destination. so that VANET system must be quite reliable.
How to make VANET reliable? This is quite a
complicated question. But there is one thing for
378
sure that security of VANET is one of necessar- Hubaux, J. P., Capkun, S., & Luo, J. (2004). The
ies to construct a reliable VANET system. Before Security and Privacy of Smart Vehicles. Security &
we design a secure system, we must find out all Privacy, 2(3), 49–55. doi:10.1109/MSP.2004.26
possible attacks first. That is what this chapter
Lin, X., Sun, X., Ho, P. H., & Shen, X. (2007).
shows us.
GSIS: A Secure and Privacy Preserving Protocol
for Vehicular Communications. IEEE Transac-
tions on Vehicular Technology, 56(6), 3442–3456.
reFerences
doi:10.1109/TVT.2007.906878
Aijz, A., Bochow, B., Dötzer, F., Festag, A., Mauve, M., Widmer, A., & Hartenstein, H. (2001).
Gerlach, M., Kroh, R., & Leinmüller, T. (2006). A Survey on Position-Based Routing in Mobile
Attacks on Inter Vehicle Communication Systems Ad Hoc Networks. IEEE Network, 15(6), 30–39.
– an Analysis. In Proc. WIT. doi:10.1109/65.967595
Armknecht, F., Festag, A., Westhoff, D., & Zeng, Raya, M., & Hubaux, J. P. (2005). The Security
K. (2007). Cross-layer Privacy Enhancement and of Vehicular Ad Hoc Networks. In Proceedings
Non-repudiation in Vehicular Communication. of the 3rd ACM workshop on Security of ad hoc
In 4th Workshop on Mobile Ad-Hoc Networks and sensor networks, (pp.11-21).
(WMAN).
Studer, A., Luk, M., Perrig, A, (n.d.). Efficient
Eichler, S., Dotzer, F., Schwingenscholgl, C., Mechanisms to Provide Convoy Member and
Javier, F., Caro, F., & Eberspacher, J, (n.d.). Secure Vehicle Sequence Authentication in VANETs.
Routing in a Vehicular Ad Hoc Network.
Yang, H., Luo, H., Ye, F., Lu, S., & Zhang, L.
Golle, P., Greene, D., & Staddon, J. (2004). Detect- (2004). Security in mobile ad hoc networks:
ing and Correcting Malicious Data in VANETs. In Challenges and solutions. IEEE Wireless com-
Proceeding of the 1st ACM international workshop munications, 11(1), 38-47.
on Vehicular ad hoc networks, Session: Security
in VANET, (pp.27-37).
Harsch, C., Festag, A., & Papadimitratos, P. (2007).
Secure Position-Based Routing for VANETs. In
IEEE 66th Vehicular Techonology Conference,
(pp. 26-30).
379
380
Compilation of References
3GPP. (2008). IP Multimedia (IM) Session Handling; IM Amit Kumar Saha, D. B. J. (2004). Modeling mobility for
Call Model; Stage 2. TS 23.218 v8.3.0. vehicular ad-hoc networks. In the 1st ACM international
workshop on Vehicular ad hoc networks (pp. 91–92). New
Aarnet - australia’s research and education network. (2009).
York: ACM Press.
Retrieved from http://www.aarnet.edu.au/
Amouris, K. (2005). Position-Based Broadcast TDMA
Aguayo, D., Bicket, J., Biswas, S., Judd, G., & Morris, R.
Scheduling for Mobile Ad-Hoc Networks (MANETs)
(2004). Link-level measurements from an 802.11b.
with Advantaged Nodes. IEEE Military Communications
Ai, Y., Sun, Y., Huang, W., & Qiao, X. (2007). OSGi based Conference, 1, 252–257.
integrated service platform for automotive telematics. In
Armknecht, F., Festag, A., Westhoff, D., & Zeng, K. (2007).
Proceedings of the IEEE International Conference on
Cross-layer Privacy Enhancement and Non-repudiation in
Vehicular Electronic and Safety, (pp. 1-6).
Vehicular Communication. In 4th Workshop on Mobile
Aidarous, S., & Plevyak, T. (1994). Telecommunications Ad-Hoc Networks (WMAN).
Network Management into the 21st Century. New York:
Artimy, M. (2007). Local Density Estimation and Dynamic
IEEE Press.
Transmission-Range Assignment in Vehicular Ad hoc Net-
Aijz, A., Bochow, B., Dötzer, F., Festag, A., Gerlach, M., works. IEEE Transactions on ITS, 18(3), 400–412.
Kroh, R., & Leinmüller, T. (2006). Attacks on Inter Vehicle
Asokan, N., Kostiainen, K., Ginzboorg, P., Ott, J., & Luo,
Communication Systems – an Analysis. In Proc. WIT.
C. (2007). Applicability of identity-based cryptography
Akingbehin, K. (2005). Wireless communications for intra- for disruption-tolerant networking. In . Proceedings of
vehicle use (Wireless harnesses). Institute for Advanced MobiOpp, 07, 52–56. doi:10.1145/1247694.1247705
Vehicle Systems, University of Michigan-Dearborn,
Assisted, G. P. S. (n.d.). Retrieved September 11, 2009, from
Dearborn, MI.
http://en.wikipedia.org/wiki/Assisted_GPS.
Akyildiz, I. F., Ho, J., & Wang, W. (1999). Mobility Manage-
ASTM International E2213-03 (2003). Standard Specification
ment in Next-Generation Wireless Systems . Proceedings
for Telecommunications and Information Exchange Between
of the IEEE, 87(8), 1347–1384. doi:10.1109/5.775420
Roadside and Vehicle Systems 5 GHz Band Dedicated Short
Alexopoulos, C. (1997). State Space Partitioning Methods Range Communications (DSRC) Medium Access Control
for Stochastic Shortest Path Problems. Networks, 1(30), (MAC) and Physical Layer (PHY) Specifications.
9–21. doi:10.1002/(SICI)1097-0037(199708)30:1<9::AID-
B. A. mobility scenario generation and analysis tool, http://
NET2>3.0.CO;2-H
web.informatik.uni-bonn.de/iv/mitarbeiter/dewaal/bon-
Alliance, O. M. (OMA). (2009). Retrieved from http://www. nmotion/.
openmobilealliance.org/
Bachir, A., & Benslimane, A. (2003). A Multicast Protocol in
AMI-C. (2002). Software API Specifications-CORE APIs. Ad hoc Networks Inter-Vehicle Geocast. IEEE Semiannual
retrieved September 20 2009, from http://www.ami-c.org Vehicular Technology Conference, 4, 2456-2460.
Copyright © 2010, IGI Global, distributing in print or electronic forms without written permission of IGI Global is prohibited.
Backes, P., Tso, K., Norris, J., Tharp, G., Slostad, J., Bonitz, Bjic, E., & Chaxel, F. (2002). Auto-ID Mobile Information
R., & All, K. (2000). Internet based Operations for the Mars System for Vehicle Life Cycle Data Management. IEEE
Polar Lander Mission. In The 2000 IEEE International Systems . Man and Cybernetics, 4, 6.
Conference on Robotics & Automation, (pp. 2025-2032).
Blaževi’c, L., Giordano, S., & LeBoudec, J. Y. (2002). Self
Bai, F., Sadagopan, N., & Helmy, A. (2003). Important: a Organized Terminode Routing. Cluster Computing Journal,
framework to systematically analyze the impact of mobility 5(2), 205–208. doi:10.1023/A:1013998030317
on performance of routing protocols for ad hoc networks.
Blum, J., Eskandarian, A., & Hoffman, L. (2004). Challenges
In The 22th IEEE Annual Joint Conference on Computer
of inter-vehicle ad hoc networks. IEEE Transactions on Intel-
Communications and Networking INFOCOM’03, (pp.
ligent Transportation Systems, 5(4), 347–351. doi:10.1109/
825–835).
TITS.2004.838218
Balasubramanian, A., Zhou, Y., Croft, B. W., Levine, B.
Bojarski, A., & Fichte, W. (2003). Temperature sensors. In J.
N., & Venkataramani, A. (2007). Web search from a bus.
Marek, H.P. Trah, Y. Suzuki,& I. Yokomori, (Ed.), Sensors
In Proceedings of the second workshop on Challenged
Application Vol. 4 - Sensors for Automotive Technology (pp.
networks CHANTS CHANTS ’07, (pp. 59–66).
343-359). Berlin: Wiley-VCH.
Baldwin, C. L. (2002). Designing in-vehicle technologies
Boneh, D., & Franklin, M. (2003). Identity based encryp-
for older drivers: application of sensory-cognitive interac-
tion from the weil pairing. In SIAM Journal of Computing,
tion theory. Theoretical Issues in Ergonomics Science, 3(4).
586–615.
doi:10.1080/1463922021000009029
Bonnick, A. (2001). Automotive computer controlled sys-
Banerjee, N., Corner, M. D., & Levine, B. N. (2007). An
tems. London: Butterworth Heinemann.
energy-efficient architecture for dtn throwboxes. In . Pro-
ceedings of IEEE INFOCOM, 2007, 776–784. Bononi, L., & Felice, M. D. (2007). A Cross Layered MAC
and Clustering Scheme for Efficient Broadcast in VANETs.
Bertsekas, D. P., & Tsitsiklis, J. N. (1991). An Analysis of Sto-
In Proceedings of the IEEE International Conference on
chastic Shortest Path Problems. Mathematics of Operations
Mobile Ad hoc and Sensor Systems.
Research, 16(3), 580–595. doi:10.1287/moor.16.3.580
Boudec, J.-Y. L., & Vojnovic, M. (2005). Perfect simulation
Bettstetter, C. (2001). Smooth is better than sharp: a random
and stationarity of a class of mobility models. In The 24th
mobility model for simulation of wireless networks. In Pro-
IEEE Annual Joint Conference on Computer Communica-
ceedings of the 4th ACM international workshop on Model-
tions and Networking INFOCOM’05, (pp. 2743–2754).
ing, analysis and simulation of wireless and mobile systems
(MSWIM ’01) (pp. 19–27). New York: ACM Press. Boyce, D. (2002). A Memoir of the ADVANCE Project.
Journal of Intelligent Transportation Systems, 7(2).
Bettstetter, C., Resta, G., & Santi, P. (2003). The node distri-
bution of the random waypoint mobility model for wireless Boyce, D. E., Kirson, A. M., & Schofer, J. L. (1994).
ad hoc networks. IEEE Transactions on Mobile Computing, ADVANCE-The Illinois Dynamic Navigation and Route
2(3), 257–269. doi:10.1109/TMC.2003.1233531 Guidance Demonstration Program. Advance Technology
for Road Transport, IVHS and ATT.
Bickel, G. S. (2006). Inter/intra-vehicle wireless commu-
nication. Retrieved from http://userfs.cec.wustl.edu/~gsb1/ Brakatsoulas, S., Pfoser, D., Salas, R., & Wenk, C. (2005).
index.html On Map-Matching Vehicle Tracking Data. In The 31st VLDB
Conference, (pp. 853 – 864).
Bicket, J., Aguayo, D., Biswas, S., & Morris, R. (2005). Ar-
chitecture and evaluation of an unplanned 802.11bmesh net- Brewer, E., Demmer, M., Du, B., Ho, M., Kam, M., &
work. In Proceedings of the 11th annual international con- Nedevschi, S. (2005). The case for technology in developing
ference on Mobile computing and networking(MOBICOM), regions. IEEE Computer, 38, 25–38.
Cologne, Germany. Available from http://www.pdos.lcs.mit.
edu/papers/roofnet: mobicom05/roofnet-mobicom05.pdf
381
Briesemeister, L., & Hommel, G. (2000). Overcoming ing. Technical report, University of Cambridge Computer
Fragmentation in Mobile Ad hoc Networks. Journal of Laboratory, Cambridge, UK.
Communications and Networks, 2(3), 182–187.
Chakravorty, R., & Pratt, I. (2002). Performance issueswith
Broggi, A., Bertozzi, M., Fascioli, A., Guarino, C., Bianco, general packet radio service. Journal of Communications
L., & Piazzi, A. (1999). The Argo Autonomous Vehicle’s and Networks (JCN), 4(2). Available from http://www.cl.cam.
Vision And Control Systems. International Journal of ac.uk/Research/SRG/netos/papers/comob-web/2002-jcn.
Intelligent Control and Systems, 3, 409–441. pdf
Brown, A., Kolberg, M., Bushmitch, D., Lomako, G., & Chan, Alex Y. M., & Lu, W. P. (2003). Architecture for
Ma, M. (2006). A SIP-based OSGi Device Communication Wireless Access in Vehicles. IEEE Vehicular Technology
Service for Mobile Personal Area Networks. In Consumer Conference, 5, 3336-3340.
Communications and Networking Conference(CCNC).
Chang, B.-J., Huang, B.-J., & Liang, Y.-H. (2008). Wireless
Buede, D. M. (2000). The engineering and design of systems: Sensor Network-based Adaptive Vehicle Navigation in
Models and methods. New York: Wiley, Inc. Multihop-Relay WiMAX Networks. The 22nd International
Conference on Advanced Information Networking and Ap-
Burgess, J., Gallagher, B., Jensen, D., & Levine, B. N.
plications, (pp. 56-63).
(2006). Maxprop: Routing for vehicle-based disruption-
tolerant networks. In . Proceedings of IEEE INFOCOM, Chen, A., Jain, N., Perinola, A., Pietraszek, T., Rooney, S.,
2006, 1–11. doi:10.1109/INFOCOM.2006.228 & Scotton, P. (2004). Scaling Real-Time Telematics Applica-
tions using Programmable Middleboxes: A Case Study in
Burns, B., Brock, O., & Levine, B. N. (2005). Mv routing
Traffic Prediction. In First IEEE Consumer Communications
and capacity building in disruption tolerant networks.
and Networking Conference, (pp. 388-393).
Proceedings of IEEE INFOCOM, 1, 398–408.
Chen, J. (1998). A Study of Web-Based SNMP Network
Bushmitch, D. Lin, Wanrong., Bieszczad, A., Kaplan, A.,
Management with a Simple Java Applet Network Monitor-
Papageorgiou, V., & Pakstas, A. (2004). A SIP-based device
ing Tool. Department of Computer Science and Engineering
communication service for OSGi framework. IEEE Consum-
Auburn University, Alabama.
er Communications and Networking Conference2006.
Chen, J. L., Chang, Y. C., Lin, Y. S., & Du, H. W. (2009).
Caliskan, M., Graupner, D., & Mauve, M. (2006). Decen-
Embedded WiMAX-based Vehicular Router for Telemat-
tralized discovery of free parking places. In Proceedings
ics Computing. In The 11th International Conference on
of the 3rd International ACM Workshop on Vehicular ad
Advanced Communication Technology, (pp. 1857-1862).
hoc networks, VANET’06, (pp. 30–39). New York: ACM
Press. Chen, Y. S., Lin, Y. W., & Lee, S. L. (2009). A Mobicast
Routing Protocol for Vehicular Ad Hoc Networks. Mobile
Camp, J., Robinson, J., Steger, S., & Knightly, E. (2006).
Networks and Applications (MONET).New York: ACM/
Measurement driven deployment of a two-tier urban mesh
Springer.
access network. In Proceedings of ACM MobiSys 2006,
Uppsala, Sweden. Available from http://networks.rice.edu/ Chen, Y. S., Lin, Y. W., & Pan, C. Y. (2009). A Diagonal-
papers/sys7122-camp.pdf Intersection-Based Routing Protocol for Urban Vehicular
Ad Hoc Networks. Telecommunication System, (under
Camp, T., Boleng, J., & Davies, V. (2002). A survey of
revision).
mobility models for ad hoc network research. Wireless
Communications and Mobile Computing (WCMC): Special Chen, Y., Zhao, W., Ammar, M., & Zegura, E. (2007). Hybrid
issue on Mobile Ad Hoc Networking: Research . Trends and routing in clustered dtns with message ferrying. In Proceed-
Applications, 2(5), 483–502. ings of the 1st international MobiSys workshop on Mobile
opportunistic networking MobiOpp ’07, (pp. 75–82).
Chaintreau, A., Hui, P., Crowcroft, J., Diot, C., Gass, R.,
& Scott, J. (2005). Pocket switched networks: Real-world Chisalita, I., & Shahmehri, E. (2004). Vehicular communi-
mobility and its consequences for opportunistic forward- cation: A candidate technology for traffic safety. In IEEE
382
International Conference on Systems, Man and Cybernetics, Devarapalli, V., Wakikawa, R., Petrescu, A., Thubert, P.
(pp. 3903–3908). (2005). Network Mobility (NEMO) Basic Support Protocol
RFC 3963.
Choffnes, D. R., & Bustamante, F. E. (2006). An integrated
mobility and traffic model for vehicular wireless networks. Differential, G. P. S. (n.d.). Retrieved September 11, 2009,
In The 2nd ACM international workshop on Vehicular ad from http://en.wikipedia.org/wiki/Differential_GPS.
hoc networks, VANET’05, (pp. 69–78). New York: ACM
Dijkstra, E. W. (1959). A note on two problems in con-
Press.
nection with graphs. Nuerische Mathematik, 1, 269–271.
Choi, D. H., Ro, S. H., Lee, K. S., Lee, W. N., Choi, S. Y., doi:10.1007/BF01386390
& Pae, Y. W. (2006). Telematics Applications Based On
Dikajakos, M. D., Florides, A., Nadeem, T., & Iftode, L.
TOPAZ Platform. In IEEE 6th International Conference on
(2008). Location-Aware Services over Vehicular Ad-Hoc
Computer and Information Technology, (pp.262-268).
Networks using Car-to-Car Communications. IEEE Journal
Choi, J. W., Han, W. Y., Kim, C. S., & Kwon, O. C. (2005). on Selected Areas in Communications, 25(8), 1590–1602.
Open Telematics Services Deployment on the Gateway and doi:10.1109/JSAC.2007.071008
Framework Independent on Mobile Networks. In Interna-
DLNA. (2009). Retrieved from http://www.dlna.org/
tional Conference on Wireless Networks, (pp.374-379).
home
Chuah, M., & Xi, Y. (2007). An encounter-based multicast
Draft 7.0, IEEE Standard 1609.1 (2006). IEEE Trial-Use
scheme for disruption tolerant networks. Technical report
Standard for Wireless Access in Vehicular Environments
published by the University of Lehigh, Lehigh, PA.
(WAVE) – Resource Manager.
Connexis website. (n.d.). Retrieved September 20, 2009,
Drane, C., & Rizos, C. (1998). Positioning system in intel-
from http://www.connexis.com/
ligent transport systems. Boston: Artech House.
CORSIM. (2009). http://www.fhwa-tsis.com/.
Dreyfus, S. E. (1969). An Appraisal of Some Shortest Path
CORSIM. http://mctrans.ce.ufl.edu/featured/tsis/version5/ Algorithms. Operations Research, 17, 395–412. doi:10.1287/
corsim.htm. opre.17.3.395
Damn small linux not (dsl-n), (n.d.). Retrieved from http:// Duri, S., Gruteser, M., Liu, X., Moskowitz, P., Perez, R.,
www.damnsmalllinux.org/dsl-n/ Singh, M., & Tang, J. M. (2002). Framework for security
and privacy in automotive telematics. In Mobile Computing
Das, S. R., Perkins, C. E., & Royer, E. M. (2000) “Perfor-
and Networking, (pp. 25-32).
mance comparison of two on-demand routing protocols for
ad hoc networks,” in Proc. IEEE Infocom. Eichler, S. (2007). Performance evaluation of the IEEE
802.11p WAVE communication standard. In Proceedings of
David, B., & David, A. (1996). Dynamic source routing in
IEEE Vehicular Technology Conference, (pp. 2199-2203).
ad hoc wireless networks. In Mobile Computing, (vol. 35,
pp. 153–181). Kluwer Academic. Eichler, S., Dotzer, F., Schwingenscholgl, C., Javier, F., Caro,
F., & Eberspacher, J, (n.d.). Secure Routing in a Vehicular
Deflorio, F. P. (2003). Evaluation of a reactive dynamic route
Ad Hoc Network.
guidance strategy. Journal of the Transportation Research
Board Transportation, 11(5), 375–388. EI-Rabbany. A. (2002). Introduction to GPS – The Global
Positioning System. Boston: Artech House.
Demers, A. A., Greene, D., Hauser, C., Irish, W., Larson, J.,
Shenker, S., et al. (1987). Epidemic algorithms for replicated Emst, T. (2006). Network Mobility Support Goals and
database maintenance. In Proceedings of the Sixth Sympo- Requirements.
sium on Principles of Distributed Computing, (pp. 1–12).
Encyclopedia.com. (2009). Telematics is not a question of
Demmer, M., & Ott, J. (2006). Delay tolerant networking if, but when. Retrieved from http://www.encyclopedia.com/
tcp convergence layer protocol. Internet-draft. doc/1G1-94875067.html
383
Ernst, T., Uehara, K., & Mitsuya, K. (2003). Network Mobil- munications and navigation system for multiple platforms.
ity from the InternetCAR Perspective. In 17th International In Proceedings of the IEEE/MTS OCEANS Conference,
Conference on Advanced Information Networking and Washington DC, (Vol. 2, pp. 1086–1092).
Applications, (pp. 19–26).
Friesz, T. L., Luque, J., Tobin, R. L., & Byung-Wook, W.
ETSI TISPAN. (2008). Retrieved from http://www.etsi. (1989). Dynamic network traffic assignment considered
org/tispan/ as a continuous time optimal control problem. Operations
Research, 37(6), 893–901. doi:10.1287/opre.37.6.893
Fall, K. (2003). A delay-tolerant network architecture for
challenged internets. In Proceedings of ACM SIGCOMM, Frost & Sullivan, (2005). Strategic Analysis of the Taiwan
(pp. 27–34). Telematics and Infotainment Market.
Farrell, J. A. (2008). Aided Navigation: GPS With High Fu, L. (2001). Adaptive Routing Algorithm for In-Vehicle
Rate Sensors (pp. 22-80). New York: Glencoe/McGraw- Route Guidance Systems with Real-Time Information.
Hill School Pub Co. Transportation Research Part B: Methodological, 35(8),
749–765. doi:10.1016/S0191-2615(00)00019-9
Farrell, S., & Cahill, V. (2005). Ltp-t: A generic delay
tolerant transport protocol. Technical report, University Fukuhara, T., Warabino, T., Ohseki, T., Saito, K., Sugiyama,
of Dublin, Dublin. K., Nishida, T., & Eguchi, K. (2005). Broadcast Methods
for Inter-Vehicle Communications System. In Proceedings
Farrell, S., Ramadas, M., & Burleigh, S. (2007). Licklider
of the IEEE Wireless Communications and Networking
transmission protocol - extensions. Internet-draft.
Conference.
Fasbender, A., Gerdes, M., Hjelm, J., Kvarnström, B., Peters-
Ganguly, S., Navda, V., Kim, K., Kashyap, K., Niculescu,
son, J., & Skog, R. (2008). Virtually at home: High-perfor-
D., Izmailov, R., et al. (2006). Performance optimizations
mance access to personal media. Ericsson Review, 2.
for deploying voip services in mesh networks. Retrieved
Feit, S. (1993). TCP/IP: Architecture, protocols and imple- from http://www.wings.cs.sunysb.edu/%7Eanand/papers/
mentation. New York: McGraw Hill, Inc. jsac06.pdf
Feit, S. (1995). SNMP: A Guide To Network Management. Geer, D. (2005). Users Make a Beeline for ZigBee Sensor
New York: McGraw Hill, Inc. Technology. Computer, 16–19.
Feit, S. (1995). SNMP: A guide to network management. Gerbers, J. (2003). High-pressure sensors. In J. Marek, H.P.
New York: McGraw Hill, Inc. Trah, Y. Suzuki,& I. Yokomori, (Ed.), Sensors Application
Vol. 4 - Sensors for Automotive Technology (pp. 333-342).
Fleming, W. J. (2008). New automotive sensors−A review. Berlin: Wiley-VCH.
IEEE Sensors Journal, 8(11), 1900–1921. doi:10.1109/
JSEN.2008.2006452 Gerlach, M., & Fokus, F. (2007). Trust for Vehicular Ap-
plications. In 8th International Symposium on Autonomous
FlexRay Communications System Protocol Specification Decentralized Systems, (pp. 295-304).
v2.1 Revision A. (2005).
Gerten, G. (2005). Protecting the Global Positioning System.
FlexRay Consortium. (2009). Retrieved from http://www. IEEE Aerospace and Electronic Systems Magazine, 20(11),
flexray.com 3–8. doi:10.1109/MAES.2005.1576067
Floyd, S., & Paxson, V. (2001) “Difficulties in simulating Global Positioning System. (n.d.). Retrieved September 11,
the Internet,” ACM/IEEE Transactions on Networking, vol. 2009, from http://en.wikipedia.org/wiki/Global_Position-
9, no. 4 (pp. 392–403). [Online]. Available: http://www.icir. ing_System.
org/vern/papers.html
Goel, A., Ramakrishnan, K. G., Kataria, D., & Logothetis,
Freitag, L., Grund, M., Singh, S., Partan, J., Koski, P., & D. (2001). Efficient Computation of Delay-Sensitive Routes
Ball, K. (2005). The whoi micro-modem: An acoustic com- from One Source to All Destinations. IEEE Conference on
Computer Communications (INFOCOM), (pp. 854–858).
384
Golle, P., Greene, D., & Staddon, J. (2004). Detecting and Radio Communications. [Online]. Available: http://jist.ece.
Correcting Malicious Data in VANETs. In Proceeding of cornell.edu/docs/050911-density.pdf
the 1st ACM international workshop on Vehicular ad hoc
Hackbarth, K. (2003). OSGi – Service-Delivery-Platform
networks, Session: Security in VANET, (pp.27-37).
for Car Telematics and Infotainment Systems. Advanced
Gong, Y., Xiong, Y., Zhang, Q., Zhang, Z., Wang, W., & Microsystems for Automotive Applications 2003 (pp. 497-
Xu, Z. (2006). Anycast routing in delay tolerant networks. 507). Berlin: Springer.
In IEEE GLOBECOM (pp. 1–5).
Hall, R. (1986). The Fastest Path through a Network with
Gorgorin, C., Gradinescu, V., Diaconescu, R., Cristea, V., Random Time- Dependent Travel Time. Transportation
& Ifode, L. (2006). An integrated vehicular and network Science, 20(3), 182–188. doi:10.1287/trsc.20.3.182
simulator for vehicular ad-hoc networks. In the 20th Euro-
Han, W. Y., Kwon, O. C., Park, J. H., & Kang, J. H. (2005).
pean Simulation and Modelling Conference.
A Gateway and Framework for Interoperable Telematics
Green, P., Levison, W., Paelke, G., & Serafin, C. (1995). Systems Independent on Mobile Networks. ETRI Journal,
Preliminary Human Factors Design Guidenlines for Driver 27(1), 106–109. doi:10.4218/etrij.05.0204.0038
Information Systems. Technical Report No. UMTRI-93-21,
Harsch, C., Festag, A., & Papadimitratos, P. (2007). Secure
Transportation Research Institute, The University of Michi-
Position-Based Routing for VANETs. In IEEE 66th Vehicular
gan, Ann Arbor, MI.
Techonology Conference, (pp. 26-30).
Grossglauser, M., & Tse, D. N. C. (2002). Mobility increases
Heidemann, J., Bulusu, N., Elson, J., Intanagonwiwat, C.,
the capacity of ad-hoc wireless networks. In IEEE/ACM
Lan, K. C., Xu, Y., et al. (2001) “Effects of detail in wireless
Transactions on Networking, 10(4), 477–486.
network simulation,” in Proc. of Communication Networks
Grymek, L., Singh, S., & Pattipati, K. (2007). Vehicular and Distributed Systems Modeling and Simulation Confer-
Dependence Adds to Telematics’ Allure. IEEE Potentials, ence, Jan. 2001. [Online]. Available: http://www.isi.edu/
26(2), 12–16. doi:10.1109/MP.2007.343053 johnh/PAPERS/Heidemann00d.html
Gsmworld.com. (2009). Retrieved from http://www.gsm- Holland, G., & Vaidya, N. H. (1999) “Proc. mobicom, acm
world.com/technology/gprs/index.shtml intern. conf. on mobile computing and networking,” in
Proc. IEEE Infocom.
Gun, N., Held, A., & Kaiser, J. (2001). Proactive services
in a distributed traffic telematics application. International Home Gateway Initiative. (2007). Retrieved from http://
Workshop on Mobile Communication over Wireless LAN: www.homegatewayinitiative.org/
Research and Applications.
Hong, X., Gerla, M., Pei, G., & Chiang, C.-C. (1999) “A
Guo, J. H., & Xing, G. M. (2007). Using Mobile Agent- group mobility model for ad hoc wireless networks,” in
based Middleware to Support Distributed Coordination Proc. ACM MSWiM.
for Vehicle Telematics. In 21st International Conference
Hong, X., Kwon, T., Gerla, M., Gu, D., & Pei, G. (2001,
on Advanced Information Networking and Applications
January) “A mobility framework for ad hoc wireless net-
Workshops, (Vol. 2, pp.374-379).
works,” in Proc. of ACM Second International Conference
Gura, N., Held, A., & Kaiser, J. (2001). Proactive Services on Mobile Data Management (MDM 2001).
in a Distributed Traffic Telematics Application. Workshop
Hsu, R. C., & Chen, L. R. (2005). Integrated Embedded
on Mobile communication over Wireless LAN: Research
System Architecture for In-Vehicle Telematics and Info-
and Applications. Retrieved from http://www.informatik.
tainment System. Proceedings of the IEEE International
uni-ulm.de/rs/projekte/core/
Symposium on Industrial Electronics, 4, 1409–1414.
Haas, Z. J., & Barr, R. (2005) “Density-independent, scal-
Hubaux, J. P., Capkun, S., & Luo, J. (2004). The Security
able search in ad hoc networks,” in Proceedings of IEEE
and Privacy of Smart Vehicles. Security & Privacy, 2(3),
International Symposium on Personal Indoor and Mobile
49–55. doi:10.1109/MSP.2004.26
385
Hughes, L., & Maghsoudlou, A. (2006). An Efficient ment 2: Physical and medium access control layers for
Coverage-Based Flooding Scheme for Geocasting in Mobile combined fixed and mobile operation in licensed bands
Ad Hoc Networks. International Conference on Advanced and corrigendum 1.
Information Networking and Applications, 1, 6.
IEEEP802.11p (2009, May). IEEE Draft Standard for Infor-
Hyyrylainen, T., Karkkainen, T., & Luo, C. (2007). Op- mation Technology – Telecommunications and information
portunistic email distribution and access in challenged exchange between systems – Local and metropolitan area
heterogeneous environments. In Proceedings of the second networks – Specific requirements – Part 11: Wireless LAN
workshop on Challenged networks CHANTS (pp.97–100). Medium Access Control (MAC) and Physical Layer (PHY)
specifications Amendment 7: Wireless Access in Vehicular
Ibm.com. (n.d.). Focus on the road ahead: IBM puts
Environments.
practical telematics within reach. Retrieved from http://
www-306.ibm.com/software/pervasive/info/Telemat- IEEEStandard1609.2 (2006). IEEE Trial-Use Standard
ics_within_reach_050404.pdf for Wireless Access in Vehicular Environments (WAVE)
– Security Services for Applications and Management
IEEE P802.11p/D3.0. (2007). Draft amendment for wireless
Messages.
access in vehicular environments (WAVE).
IETF Network Working Group. (2008). RFC3344- IP Mo-
IEEE Standard 1609.3 (2007). IEEE Trial-Use Standard
bility Support for IPv4. Retrieved from http://www.faqs.
for Wireless Access in Vehicular Environments (WAVE) –
org/ rfcs/rfc3344.html
Networking Services.
InCode Telecom Group Inc. (2009). Telematics: How eco-
IEEE Standard 1609.4 (2006). IEEE Trial-Use Standard
nomic and technological forces will shape the industry in
for Wireless Access in Vehicular Environments (WAVE) –
the U.S. Retrieved from http://www.incodewireless.com/
Multi-channel Operation.
pdfMailer/files/Telematics_Position_Paper_v11.pdf
IEEE Std 1488-1999 (2000). IEEE Standard for Message
Institute of Transportation Engineers Management and
Set Template for Intelligent Transportation Systems.
Operations of Intelligent Transportation Systems. (2003).
IEEE Std 1489-1999 (1999). IEEE Standard for Data Dic- ITS standards overview.
tionaries for Intelligent Transportation Systems.
Internet Librarian, D. I. S. A. (1997). US-DOD Internet
IEEE Std 828-1998. (1998). IEEE standard for software related standardized profiles. Retrieved from http://www-
configuration management plans. library .itsi.disa.mil/
IEEE1609.1TM (2006). IEEE Trial-Use Standard for Wireless Internet, C. A. R. Project (2003). InternetCAR webpage at
Access in Vehicular Environments (WAVE) – Resource WIDE. Retrieved September 20, 2009 from http://www.sfc.
Manager. wide.ad.jp/InternetCAR
IEEE1609.2TM (2006). IEEE Trial-Use Standard for Wire- Interplanetary internet project (n.d.). Interplanetary internet
less Access in Vehicular Environments (WAVE) – Security project. Retrieved from http://www.ipnsig.org
Services for Applications and Management Message.
Ip, Y. K., Lau, W. C., & Yue, O. C. (2007). Forwarding and
IEEE1609.3 (2006). IEEE Trial-Use Standard for Wireless
TM replication strategies for dtn with resource constraints. In
Access in Vehicular Environments (WAVE) – Networking . Proceedings of IEEE Vehicular Technology Conference,
Services. 1, 1260–1264.
IEEE1609.4TM (2006). IEEE Trial-Use Standard for Wire- ITS Standards Outreach, Education and Training Program,
less Access in Vehicular Environments (WAVE) – Multi- Institute of Transportation Engineers Management and
channel Operation. Operations of Intelligent Transportation Systems (2003).
ITS Standards Overview.
IEEE802.16-2005. (2006). Part 16: Air interface for fixed
and mobile broadband wireless access systems amend-
386
ITS Standards Outreach, Education and Training Program, Joshi, H. P., Sichitiu, M., & Kihl, M. (2007). Distributed
Institute of Transportation Engineers. (2006). Center to Robust Geocast Multicast Routing for Inter-Vehicle Com-
center communications. munication. WEIRD Workshop on WiMax, Wireless and
Mobility.
ITU-T X.680-X.690 (1994). ISO/IEC 8824: Abstract Syntax
Notation One (ASN.1). Jost, D., & Nagel, K. (2003). Probabilistic traffic flow break-
down in stochastic car following models. Transportation
Jain, S., Fall, K., & Patra, R. (2004). Routing in a delay
Research Record, 1852, 152–158. doi:10.3141/1852-19
tolerant network. In Proceedings of ACM SIGCOMM, (Vol.
34, pp. 145–158). Jun, H., Ammar, M. H., & Zegura, E. W. (2005). Power
management in delay tolerant networks: A framework and
Jakoby, B., & Herrmann, F. (2003). Chemical sensors for liq-
knowledge-based mechanisms. In IEEE Communications
uid media. In J. Marek, H.P. Trah, Y. Suzuki,& I. Yokomori,
Society Conference on Sensor and Ad Hoc Communications
(Ed.), Sensors Application Vol. 4 - Sensors for Automotive
and Networks (SECON), (pp. 418–429, 2005).
Technology (pp. 516-526). Berlin: Wiley-VCH.
Jun, H., Ammar, M. H., Corner, M. D., & Zegura, E. W.
Jardosh, A. P., Belding-Royer, E. M., Almeroth, K. C., &
(2006). Hierarchical power management in disruption toler-
Suri, S. (2005) “Real-world environment models for mobile
ant networks with traffic-aware optimization. In Proc. ACM
network evaluation,” in In the IEEE Journal on Special
SIGCOMM Workshop on Challenged Networks (CHANTS),
Areas in Communications - Special Issue on Wireless Ad
(pp. 245–252).
hoc Networks.
Jung, C. R., & Kelber, C. R. (2005). Lane following and
Jardosh, A., Ramachandran, K., Almeroth, K. C., &
lane departure using a linear-parabolic model. Image
Belding-Royer, E. M. (2005). Understanding congestion
and Vision Computing, 23, 1192–1202. doi:10.1016/j.ima-
in ieee 802.11b wireless networks. In Proceeding of ACM
vis.2005.07.018
SIGCOMM Internet Measurement Conference, Berkeley,
CA. Retrieved from http://moment.cs.ucsb.edu/amitj/ Juo, C. S., & Pan, J. Y. (2008). Software Agent Framework
jardosh-imc2005.pdf for Dynamic Handoff Decision. In APCC 2008, Akihabara,
Tokyo, Japan. Ergen, M. (2009). The Access Service Network
Jiang, D., & Delgrossi, L. (2008). IEEE 802.11p: Towards
in WiMAX: The Role of ASN-GW.
an international standard for wireless access in vehicular
environments. In IEEE Vehicular Technology Conference, Kaplan, E. D., & Hegarty, C. J. (2006). Understanding GPS
(pp. 2036-2040). – Principle and Applications. Boston: Artech House.
Johansson, K. H., Torngren, M., & Nielsen. L. (2003). Vehicle Karimi, A., Hegyi, A., Schutter, B. D., Hellendoorn, J., &
applications of controller area network. Technical Report Middelham, F. (2004, Oct.). Integrated model predictive
Department of Signals, Sensors and Systems, Royal Institute control of dynamic route guidance information systems and
of Technology, Stockholm, Sweden, Department of Electri- ramp metering. Paper presented at the Proceedings of the 7th
cal Engineering, Linkoping University, Sweden. International IEEE Conference on Intelligent Transportation
Systems (ITSC 2004), Washington, DC.
Johansson, P., Larsson, T., & Hedman, N. (1999). Scenario-
based performance analysis of routing protocols for mobile Karnadi, F. K., Mo, Z. H., & Lan, K. C. (2007). Rapid
ad-hoc networks. In MOBICOM 99, The Fifth Annual ACM/ generation of realistic mobility models for vanet. In IEEE
IEEE International Conference on Mobile Computing and Wireless Communications and Networking Conference.
Networking, Seattle, WA (pp. 15–19).
Karp, B. (2001). Challenges in Geographic Routing: Sparse
Johansson, P., Larsson, T., Hedman, N., Mielczarek, B., & Networks, Obstacles, and Traffic Provisioning. In DIMACS
Degermark, M. (1999) “Routing protocols for mobile ad-hoc Workshop on Pervasive Networking.
networks - a comparative performance analysis,” in IEEE
Karp, B., & Kung, H. T. (2000). GPSR: Greedy Perimeter
MOBICOM, p. 195V206. K. tutorial, http://www.keyhole.
Stateless Routing for Wireless Networks. ACM Interna-
com/kml/kml tut.html.
tional Conference on Mobile Computing and Networking,
(pp. 243-254).
387
Kastinaki, V., Zervakis, V., & Kalaitzakis, K. (2003). A sur- wireless ad hoc networks. In IEEE Vehicular Technology
vey of video processing techniques for traffic applications. Conference (VTC), (pp. 2289– 2293).
Image and Vision Computing, 21(4), 359–381. doi:10.1016/
Lee, C., Nordstedt, D., & Helal, S. (2003). Enabling Smart
S0262-8856(03)00004-0
Spaces with OSGi. IEEE Pervasive Computing, 2(3, July-
Kim, J. H., Seo, S. H., Moon, T. Y., Hwang, S. H., & Jeon, Sept), 89-94.
J. W. (2007). A method of improving the reliability of the
Lee, J., Forlizzi, J., & Hudson, S. E. (2005). Studying the
gateway system by using OSEK/VDX. In Proceedings of
effectiveness of MOVE: a contextually optimized in-vehicle
IEEE International Conference on Control, Automation
navigation system. In Conference on Human Factors in
and System, (pp. 2328-2833).
Computing Systems.
Kim, M. J., Choi, Y. J., Moon, Y. B., Kim, S. J., & Kwon,
Lee, K. K., Kim, S. H., Choi, Y. S., & Park, H. S. (2006).
O. C. (2006). Design and Implementation of Status based
A mesh routing protocol using cluster label in the ZigBee
Application Manager for Telematics. IEEE 8th International
network. In Proceedings of Mobile Ad-hoc and Sensor
Conference on Advanced Communication Technology,
Systems (MASS), (pp. 801-806).
(pp. 20-22).
Lee, S. W., Munson, J., Lee, D. R., Thompson, G., & Park,
Kim, T., Kim, D., Park, N., Yoo, S., & Lopez, T. S. (2007).
J. (2005). A Rule-based System for Sense-and-Respond
Shortcut tree routing in ZigBee networks. In . Proceedings
Telematics Applications. In The First Korea/Japan Joint
of ISWPC, 07, 42–47.
Workshop on Ubiquitous Computing & Networking Sys-
Kofink, P. (2003). Steering-angle sensors. In J. Marek, H.P. tems, (pp 109-114).
Trah, Y. Suzuki,& I. Yokomori, (Ed.), Sensors application
Lee, W., Tak, Y., Byun, W. H., & Pae, Y. W. (2005). Toward
Vol. 4 - Sensors for automotive technology (pp. 428-449).
an Open Telematics Infrastructure. IPSJ SIG Technical
Berlin: Wiley-VCH.
Reports, (pp. 443-448).
Korkmaz, G., Ekici, E., & Ozguner, F. (2006). A Cross-
Leguay, J., Friedman, T., & Conan, V. (2006). Evaluating
Layer Multihop Data Delivery Protocol with Fairness
mobility pattern space routing for DTNs. In Proceedings
Guarantees for Vehicular Networks. IEEE Transactions
of IEEE INFOCOM.
on Vehicular Technology, 55(3), 865–875. doi:10.1109/
TVT.2006.873838 Leick, A. (2004). GPS Satellite Surveying. New York: John
Wiley & Sons.
Kotsialos, A., Papageorgiou, M., Diakaki, C., Pavlis, Y.,
& Middelham, F. (2002). Traffic flow modeling of large- Li, Y., Wang, F., Feng, H., & Li, Z. (2005). OSGi-based
scale motorway networks using the macroscopic modeling service gateway architecture for intelligent automobiles.
tool METANET. Journal of IEEE Transaction on Intel- In Proceedings of the IEEE International Conference on
ligent Transportation Systems, 3(4), 282–292. doi:10.1109/ Vehicles Symposium, (pp. 861-865).
TITS.2002.806804
Ligs, J. F., & Bowcott, S. (1995). ADVANCE: Initial Deploy-
Krotkov, E., & Blitch, J. (1999). The defense advanced ment. Annual Meeting of ITS America.
research projects agency (darpa) tactical mobile robotics
Lin, X., Sun, X., Ho, P. H., & Shen, X. (2007). GSIS: A
program. In The International Journal of Robotics Research
Secure and Privacy Preserving Protocol for Vehicular Com-
(pp. 769–776).
munications. IEEE Transactions on Vehicular Technology,
Kuhne, R. D., & Langbein-Euchner, K. (1995). Calculation 56(6), 3442–3456. doi:10.1109/TVT.2007.906878
of travel time savings by dual mode route guidance for the
Lindgren, A., & Doria, A. (2007). Experiences from deploy-
South corridor in the Stuttgart test field. Vehicle Navigation
ing a reallife dtn system. In 2007 4th IEEE Consumer Com-
and Information Systems Conference.
munications and Networking Conference, (pp. 217–221).
LeBrun, J., Chuah, C.-N., Ghosal, D., & Zhang, M. (2005).
Lindgren, A., Doria, A., & Schelen, O. (2003). Probabilistic
Knowledgebased opportunistic forwarding in vehicular
routing in intermittently connected networks. ACM SIG-
388
MOBILE Mobile Computing and Communications Review, Marfia, G., Pau, G., Giordano, E., Sena, E. D., & Gerla, M.
7(3), 19–20. doi:10.1145/961268.961272 (2007) “Evaluating vehicle network strategies for downtown
portland: Opportunistic infrastructure and importance
Lochert, C., Mauve, M., Füßler, H., & Hartenstein, H. (2005).
of realistic mobility models,” in proc. of MoBiOpp 2007,
Geographic routing in city scenarios. ACM SIGMOBILE
co-located with ACM/USENIX International Conference
Mobile Computing and Communications, 9(1), 69–72.
on Mobile Systems, Applications, and Services (MobiSys
doi:10.1145/1055959.1055970
2007).
Locust world. (2009). Retrieved from http://www.locust-
Mariño, P., Poza, F., Dominguez, M. A., & Otero, S. (2009).
world.com
Electronics in automotive engineering: A top-down ap-
Lundgren, H., Ramachandran, K., Belding-Royer, E., proach for implementing industrial field bus technologies
Almeroth, K., Benny, M., Hewatt, A., Touma, A., & Jar- in city buses and coaches. IEEE Transactions on Industrial
dosh, (2006). Experiences from building and using the Electronics, 589–600. doi:10.1109/TIE.2008.2002723
ucsb meshnet testbed. IEEE Wireless network. Available
Marples, D., & Kriens, P. (2001). The Open Services Gateway
from http://user.it.uu.se/[UNKNOWN ENTITY &tildent;]
Initiative: An Introduction Overview. IEEE Communica-
henrikl/publications.html mesh network.
tions Magazine, 39(12), 110–114. doi:10.1109/35.968820
Luo, J., & Hubaux, J.-P. (2004). A survey of inter-vehicle
Mauve, M., Widmer, A., & Hartenstein, H. (2001). A Survey
communication. School of computer and Communication
on Position-Based Routing in Mobile Ad Hoc Networks.
Sciences, EPEL, Tech. Rep. IC/2004/24.
IEEE Network, 15(6), 30–39. doi:10.1109/65.967595
Machen. L. (2009). Technical marketing engineer, “Intel
Messmer, & Papageorgiou, M. (1990). METANET: A macro-
drives in-vehicle solutions” handheld components division.
scopic simulation program for motorway networks. Journal
Retrieved from http://www.intel.com/technology/magazine/
of Traffic Engineering & Control, 31(8-9), 466-470.
computing/it04001.pdf
Michael, A. (2004). Guide to the IEEE 1512 family of
Mahajan, A., Potnis, N., Gopalan, K., & Wang, A.-I. A.
standards. Washington, DC: Institute of Electrical &
(2006). Urban mobility models for vanets. In Proceedings
Electronics Engineer.
of the 2nd IEEE International Workshop on Next Genera-
tion Wireless Networks. Microsoft drives to Las Vegas. (2004). Connected Concept
Cars. Retrieved from http://www.windowsfordevices.com/
Mahfoud, M., Al-Holou, N., & Baroody, R. (2008). Next
news/NS5155857065.html
generation vehicle network: Web enable. In Proceedings
of 3rd International Conference on Information and Com- Microsoft.com. (2009). Windows Automotive. Retrieved
munication Technologies: From Theory to Applications, from http://www.microsoft.com/windowsautomotive/wa5/
(pp. 1-7). default.mspx.
Mangharam, R., Weller, D. S., Stancil, D. D., Rajkumar, R., Minciardi, R., & Gaetani, F. (2001). A decentralized optimal
& Parikh, J. S. (2005) “GrooveSim: A topography-accurate control scheme for route guidance in urban road networks.
simulator for geographic routing in vehicular network,” in Paper presented at the Proceeding of IEEE Intelligent
Proc. of the 2nd ACM International Workshop on Vehicular Transportation Systems, Oakland, CA.
Ad Hoc Networks(VANET),Cologne,Germany.[Online].
Mobile taiwan applications promotion project m-taiwan.
Available: http://www.andrew.cmu.edu/user/rahulm/Re-
Retrieved from http://www.pwlan.org.tw/mp.asp?mp=3
search/Pubs/ vanet05 mangharam.pdf
Monk, D., Mladenovic, D., & Skaw, M. (2003). Accel-
Manoharan, R., & Thambidurai, S. L. P. P. (2008). Energy
erometers for automotive applications. In J. Marek, H.P.
Efficient Robust On-Demand Multicast Routing Protocol
for MANETs. [IJAHUC]. International Journal of Ad
Vol. 4 - Sensors for automotive technology (pp. 269-296).
Hoc and Ubiquitous Computing, 3(2), 90–98. doi:10.1504/
Berlin: Wiley-VCH.
IJAHUC.2008.017002
389
Morbe, M., & Zwiener, G. (2003). Whell-speed sensors. In J. International Conference on Computer Communications
Marek, H.P. Trah, Y. Suzuki,& I. Yokomori, (Ed.), Sensors (INFOCOM’07), (pp. 1919-1927).
application Vol. 4 - Sensors for automotive technology (pp.
Naumov, V., Baumann, R., & Gross, T. (2006). An Evaluation
Of Inter-Vehicle Ad Hoc Networks Based On Realistic Ve-
MOSQUITO Project. (2005). Mosquito Architecture. White hicular Traces. In ACM International Symposium on Mobile
Paper. Retrieved September 20, 2009 from http://www. Ad Hoc Networking and Computing, (pp. 108-119).
mosquito-online.org/
Next Generation Telematics Protocol website. (n.d.). Re-
Mosra, S. R., Shanker, S., & Mahmud, S. M. (2004). An trieved September 20 2009, from http://www.ngtp.org
intelligent architecture for issuing intersection collision
NMEA Data. (n.d.). Retrieved September 11, 2009, from
warnings (pp. 176-183). National Defense Industries As-
http://gpsinformation.org/dale/nmea.htm.
sociation (NDIA). OSEK VEX Portal, (n.d.). Retrieved
from http://www.osek-vdx.org Noble, B., Yoon, J., & Liu, M. (2007, June 11-14). Surface
Street Traffic Estimation. Paper presented at the Proceed-
Muffat, M., Green, P., & Crosnier, S. (1991). European
ings of International conference on Mobile systems, San
cooperation on dual mode route guidance—Perspectives
Juan, Puerto Rico.
for advanced research partners. Vehicle Navigation and
Information Systems Conference. PaPaGO SDK. (n.d.). Nolte, T., Hansson, H., & Bello, L. L. (2005). Automotive
Retrieved from http://www.papagosdk.com/ Communication – Past, Current and Future. In 10th IEEE
Conference on Emerging Technologies and Factory Auto-
Munson, J., Lee, S. W., Lee, D. R., Wood, D., Thompson,
mation, 1, 985-992.
G., & Cole, A. (2005). A Rule-based System for Sense-and-
Respond Telematics Services. Workshop on End-to-end, Normann, N., Schulze, G., & Keller, W. (2003). Tire-pressure
Sense-and-Respond Systems, Applications and Services, sensors. In J. Marek, H.P. Trah, Y. Suzuki,& I. Yokomori,
(pp. 31-36). (Ed.), Sensors application Vol. 4 - Sensors for automotive
technology (pp. 536-544). Berlin: Wiley-VCH.
Musolesi, M., Hailes, S., & Mascolo, C. (2005). Adaptive
routing for intermittently connected mobile ad hoc networks. Nowey, K. P. T., & Mletzko, C. (2006). Towards a Secu-
In IEEE WoWMoM 2005 (pp. 183–189). rity Architecture for Vehicular ad hoc Networks. In First
International Conference on Availability, Reliability and
Myoung, K., Heo, J., Kwon, W. H., & Kim, D. S. (2005,
Security, (pp. 374-381).
July). Design and implementation of home network control
protocol on OSGi for home automation. In . Proceedings NTCIP Standard 1103 (2005). TS 3.2-1996 NTCIP Simple
of the IEEE International Conference on Advanced Com- Transportation Management Framework - Amendment 1.
munication Technology, 2, 1163–1168.
NTCIP Standard 9001. (2002). The NTCIP guide.
Nadeem, T., Shankarand, P., & Iftode, L. (2006, July). A
O’Donnell, M. (2003). OSEK/VDX: Driving the Standard
comparative study of data dissemination models for vanets.
for Optimized Embedded Systems. Embedded Computing
Paper presented at the Proceeding of 3rd Annual Inter-
Design.
national Conference on Mobile and Ubiquitous Systems
(MOBIQUITOUS), San Jose, CA. Open Service Gateway Initiative. (2003). OSGi Specifica-
tion Release 3. Retrieved September 20 2009, from http://
Nagatani, T. (2002). The physics of traffic jams. Reports on
www.osgi.org
Progress in Physics, 65(9), 1331–1386. doi:10.1088/0034-
4885/65/9/203 OSEK Group (2004). OSEK Implementation Language,
Version 2.5.
National Highway Institute (2001). Using the National ITS
Architecture for Deployment. OSEK Group (2004). OSEK/VDX Network Management
Concept and Application Programming Interface, Ver-
Naumov, V., & Gross, T. (2007). Connectivity-Aware
sion 2.5.3.
Routing (CAR) in Vehicular Ad Hoc Networks. IEEE
390
OSEK Group (2005). OSEK/VDX Operation System, Ver- Pazul. K. (1999). An introduction to the CAN protocol that
sion 2.2.3. discusses the basics and key features. Microchip Applica-
tion Note #AN713.
Ott, J., & Kutscher, D. (2006). Bundling the web: Http over
dtn. In Proceeding of WNEPT2006. Perkins, C. E., & Royer, E. M. (1999). Ad-hoc On-demand
distance vector routing. In . Proceedings of WMCSA, 99,
P. simulation VISSIM. (2009). http://www.english.ptv.de/.
90–100.
Palazzi, C. E., Roccetti, M., Pau, G., & Gerla, M. (2007).
Perkins, D., & McGinnis, E. (1997). Understanding SNMP
Online Games on Wheels: Fast Game Event Delivery in
MIBS. Upper Saddle River, NJ: Prentice Hall, Inc.
Vehicular Ad-hoc Networks. In IEEE Intelligent Transpor-
tation Systems Society, (pp. 42–49). Phung, P. H., & Sands, D. (2008). Security Policy Enforce-
ment in the OSGi Framework Using Aspect-Oriented
Panichpapibo, S., & Pattara-atikom, W. (2008, Oct.).
Programming. In Proceedings of the IEEE International
Evaluation of a neighbor-based vehicle density estimation
Conference on Computer Software and Applications,
scheme. Paper presented at the ITST 2008. 8th International
(pp.1076-1082).
Conference, Phuket.
Popescu, D., & Selisteanu, D. (2008). Web based Telem-
Paret, D. (2007). Multiplexed networks for embedded systems
atics Application for Robotics. In the 3rd International
– CAN, LIN, FlexRay, Safe-by-Wire. Hoboken, NJ: Wiley.
Multi-Conference on Computing in the Global Information
Park, C. K., Kang, J. M., Choi, M. J., Hong, J. W., Lim, Technology, (pp. 19-24).
Y. H., & Choi, M. S. (2008, Jan.). An Integrated Network
Potnis, N., & Mahajan, A. (2006) “Mobility models for
Management System for Multi-Vendor Power Line Commu-
vehicular ad hoc network simulations,” in proc. of the
nication Networks. In Proceedings of the IEEE International
44th annual Southeast regional conference, ACM (pp.
Conference on Information Networking, (pp.23-25).
746–747).
Park, C. K., Kang, J. M., Choi, M. J., Hong, J. W., Lim,
Prosyst website. (n.d.). Retrieved September 20 2009, from
Y. H., Ju, S., & Choi, M. S. (2008). Definition of common
http://www.prosyst.com/
PLC MIB and design of MIB Mapper for multi-vendor
PLC network management. In Proceedings of the IEEE Punnoose, J., Tseng, S., Wang, S., Nikitin, V., & Schlesinger,
International Symposium on Power Line Communications T. E. (2001). Communications Resources Management for
and Its Applications, (pp.152-157). Advanced Telematics Applications. Transportation Society
Conference.
Park, J. G., Ahn, C. W., Cho, H. N., Byun, I. S., Desmons,
F., & Kim, S. W. (2006). A method for representing and Rajkumar, R., Gagliardi, M., & Sha, L. (1995). The Real-
transporting CIM operations using binary XML in the Time Publisher/Subscriber Inter-Process Communication
WBEM architecture. In Proceedings of the 8th International Model for Distributed Real-Time Systems: Design and
Conference on Advanced Communication Technology, Implementation. Real-Time Technology and Applications
(Vol. 3, pp.20-22). Symposium, (pp. 66–75).
Park, J. S., Lee, U., Oh, S. Y., Gerla, M., & Lun, D. (2006). Ramadas, M., Burleigh, S., & Farrell, S. (2007). Licklider
Emergency Related Video Streaming in VANETs using transmission protocol - specification. Internet-draft.
Network Coding. In ACM International Workshop on Ve-
Rapport, T. S. (1996). Wireless communications principles
hicular Ad Hoc Networks (VANET), (pp. 102-103).
and practice. Upper Saddle River, NJ: Prentice Hall.
Pavlis, Y., & Papageorgiou, M. (1999). Simple Decentralized
Raya, M., & Hubaux, J. P. (2005). The Security of Vehicular
Feedback Strategies for Route Guidance in Traffic Net-
Ad Hoc Networks. In Proceedings of the 3rd ACM workshop
works. Transportation Science, 33(3), 264–278. doi:10.1287/
on Security of ad hoc and sensor networks, (pp.11-21).
trsc.33.3.264
Raychaudhuri, D., Seskar, I., Ott, M., Ganu, S., Ramachan-
dran, K., Kremo, H., et al. (2005). Overview of the orbit
391
radio grid testbed for evaluation of next-generation wireless Roessler, B., & Bruns, C. (2006, Mar.). Replication of Local
network protocols. In Proceedings of the IEEE Wireless Road Networks for Regional Information Dissemination
Communications and Networking Conference, New Or- in VANETS. Paper presented at the Proceedings of the 3rd
leans, LA, USA. Workshop on positioning, Navigation And Communication
(WPNC’06), University of Hannover, Germany.
Real-Time Traffic Network Systems of Cities and Coun-
ties in Taiwan. (n.d.). Retrieved from http://www.e-traffic. Rose, M. T. (1990). The Open Book: A Practical Perspective
com.tw on OSI. Upper Saddle River, NJ: Prentice Hall, Inc.
Reilly, D., & Taleb-Bendiab, A. (2002). A Service-Based Rosen, J., Schulzrinne, H., Camarillo, G., Johnston, A.,
Architecture for In-Vehicle Telematics Systems. In the Peterson, J., Sparks, R., et al. (2002). SIP: Session Initiation
22nd International Conference on Distributed Computing Protocol. IETF RFC 3261.
Systems Workshops, (pp.741-742).
Royer, E. M., Melliar-Smithy, P. M., & Mosery, L. E. (2001)
Research and Innovative Technology Administration (RITA) “An analysis of the optimum node density for ad hoc mobile
& U.S. Department of Transportation. (US DOT). (2002). networks,” in Proc. of the IEEE International Conference
National ITS architecture. Retrieved from http://itsarch. on Communications, Jun. 2001.
iteris.com/itsarch/
Ryu, B., Yeung, G., Krishnan, H., Yin, J., ElBatt, T., Hab-
Rice, J., & Cho, Y. (2006). Estimating Velocity Fields on ermas, S., et al. (2004). Performance Evaluation of Safety
a Freeway From Low-Resolution Videos. IEEE Transac- Applications over DSRC Vehicular Ad Hoc Networks. Paper
tions on Intelligent Transportation Systems, 7(4), 463–469. presented at the Proceedings of the 1st ACM international
doi:10.1109/TITS.2006.883934 workshop on Vehicular ad hoc networks, Philadelphia,
PA.
Riegel, J., Wiedenmann, H., & Neumann, H. (2003). Chemi-
cal sensors for oxygen detection and air/fuel ratio contol. In S. S. of Urban Mobility. (2009). http://sumo.sourceforge.
J. Marek, H.P. Trah, Y. Suzuki,& I. Yokomori, (Ed.), Sensors net/.
application Vol. 4 - Sensors for automotive technology (pp.
Safa, H., Artail, H., & Shibli, R. (2009). An Interoperability
Model for Supporting Reliability and Power-Efficient Rout-
Robert Bosch Gmb, H. (2006). Safety, comfort and conve- ing in MANETs. [IJAHUC]. International Journal of Ad
nience systems. Hoboken, NJ: Wiley. Hoc and Ubiquitous Computing, 4(2), 71–83. doi:10.1504/
IJAHUC.2009.023898
Robert Bosch Gmb, H. Stuttgart, Germany, (1991). CAN
Specification ver. 2.0. Saha, A. K., & Johnson, D. B. (2004) “Modeling mobility
for vehicular ad hoc networks,” in Proc. of the 2nd ACM
Robert, H., & Ilya, P. (1979). A Two-Fluid Approach to
International Workshop on Vehicular Ad Hoc Networks
Town Traffic. Science, 204(4389), 148–151. doi:10.1126/
(VANET). [Online]. Available: http://www.cs.rice.edu/
science.204.4389.148
amsaha/ Resume/vanet2004/vanet2004.pdf
Robinson, C. L., Caveney, D., Laberteaux, K., & Caminiti, L.
Saha, A. K., & Johnson, D. B. (2004). Modeling mobility for
(2006, September 29). Efficient Coordination and Transmis-
vehicular ad hoc networks. ACM International Workshop on
sion of Data for Cooperative Vehicular Safety Applications.
Vehicular Ad Hoc Networks (VANET), (pp. 91-92).
Paper presented at the Proceedings of the 3rd international
workshop on Vehicular ad hoc networks, Los Angeles, Saito, T., Tomoda, I., Takabatake, Y., Arni, J., & Teramoto,
California, USA. K. (2000). Home gateway architecture and its implementa-
tion. IEEE Transactions on Consumer Electronics, 46(4),
Rodoplu, V., & Park, M. K. (2005). An energy-efficient
1161–1166. doi:10.1109/30.920474
mac protocol for underwater wireless acoustic networks. In
Proceedings of the MTS/IEEE OCEANS’05 Conference. Sawamura, T., Tanaka, K., Atajanov, M., Matsumoto, N., &
Yoshida, N. (2008). Adaptive Router Promotion and Group
Forming in Ad-Hoc Networks. [IJAHUC]. International
392
Journal of Ad Hoc and Ubiquitous Computing, 3(4), 217–223. Sichitiu, M. L., & Kihl, M. (2008). Inter-Vehicle Com-
doi:10.1504/IJAHUC.2008.018907 munication Systems: A Survey. IEEE Communica-
tions Surveys & Tutorials, 10(2), 88–105. doi:10.1109/
Schatz, O. (2003). Yaw-rate sensors. In J. Marek, H.P.
COMST.2008.4564481
Vol. 4 - Sensors for automotive technology (pp. 297-313). SimulationP. M. T. (2009). http://www.paramics-online.
Berlin: Wiley-VCH. com/.
Schiller, J., & Voisard, A. (2004). Location-Based Services. SimulatorO. (2009). http://www.opnet.com/.
San Francisco: Morgan Kaufmann.
SimulatorQ. N. (2009). http://www.scalable-networks.
Schilling, K. (2001). Virtual Laboratories for Mobile Robots. com/.
IFAC Workshop on Internet Based Control Education, (pp.
Singh, M., Ott, M., Seskar, I., & Kamat, P. (2005). Orbit
115-120).
measurements framework and library (oml): Motivations,
Schmitt, D. (2003). Chemical sensors for emission control. design,implementation, and features. In Proceedings of
In J. Marek, H.P. Trah, Y. Suzuki,& I. Yokomori, (Ed.), Sen- IEEE Tridentcom 2005, Trento, Italy.
sors application Vol. 4 - Sensors for automotive technology
Skordylis, A., & Trigoni, N. (2008). Delay-Bounded Rout-
(pp. 500-508). Berlin: Wiley-VCH.
ing in Vehicular Ad-Hoc Networks. ACM international
Schmitt, E. J., & Jula, H. (2006, Sep.). Vehicle Route symposium on Mobile ad hoc networking and computing,
Guidance Systems: classification and Comparison. Paper (pp. 341-350).
presented at the Proceedings of the IEEE Intelligent Trans-
Sormani, D., Turconi, G., Costa, P., Frey, D., Migliavacca,
portation Systems Conference (ITSC2006), Toronto.
M., & Mottola, L. (2006). Towards Lightweight Information
Scott, K. (2005). Disruption tolerant networking proxies Dissemination in Inter-Vehicular Networks. In Proceed-
for on-the-move tactical networks. In IEEE MILCOM 2005, ings of the ACM International Workshop on Vehicular
(Vol. 5, pp. 3226–3231). Inter-Networking.
Scott, K., & Burleigh, S. (2007). Bundle protocol specifica- South Pointing Chariot. (n.d.). Retrieved September 11,
tion. Internet- Draft. 2009, from http://en.wikipedia.org/wiki/South_Point-
ing_Chariot.
Segui, J., & Jennings, E. (2006). Delay tolerant networking
v bundle protocol simulation. In Second IEEE International Spyropoulos, T., Psounis, K., & Raghavendra, C. S. (2004).
Conference on Space Mission Challenges for Information Single-copy routing in intermittently connected mobile
Technology. networks. In 2004 First Annual IEEE Communications
Society Conference on Sensor and Ad Hoc Communications
Seo, S. H., Moon, T. Y., & Kim, J. H. (2007). Smart Vehicle
and Networks, (pp. 235 – 244).
Management System by using Gateway, Hand-set and
VMP. International Conference on Control, Automation Spyropoulos, T., Psounis, K., & Raghavendra, C. S. (2005).
and Systems, (pp.17-20). Spray and wait: An efficient routing scheme for intermit-
tently connected mobile networks. In Proceeding of the 2005
Seongmoon, K., & Lewis, M. E., & III, C. C. W. (2005).
ACM SIGCOMM workshop on Delay-tolerant networking
Optimal vehicle routing with real-time traffic information.
WDTN ’05, (pp. 252–259).
IEEE Transactions on Intelligent Transportation Systems,
6(2), 178–188. doi:10.1109/TITS.2005.848362 Srisuresh, P., Kuthan, J., Rosenberg, J., Molitor, A., & Ray-
han, A. (2002). Middlebox Communication Architecture and
Seth, A., & Keshav, S. (2005). Practical security for dis-
Framework. IETF Internet Draft, retrieved September 20
connected nodes. In Proceeding of Workshop on Secure
2009 from http://www.jdrosen.net/papers/rfc3303.txt
Network Protocols (pp. 31–36).
Stallings, W. (1993). SNMP, SNMPv2 and CMIP The Prac-
tical Guide to Network Management Standards. Reading,
MA: Addison-Wesley Publishing Company, Inc.
393
Stallings, W. (1996). SNMP, SNMPv2 and RMON. Reading, Symington, S., Farrell, S., Weiss, H., & Lovell, P. (2007).
MA: Addision-Wesley Publishing Company, Inc. Delay-tolerant networking security overview. Internet-
draft.
Steedman, D. (1993). ASN.1: The Tutorial & Reference.
London: Technology Appraisals Ltd. T. N. S. ns 2.(2009). http://www.isi.edu/nsnam/ns/index.
html.
Stepanov, I., Marron, P.-J., & Rothermel, K. (2005l) “Mobil-
ity modeling of outdoor scenarios for manets,” in Proceed- Taleb, T., Sakhaee, E., Jamalipour, A., Hashimoto, K., Kato,
ings of the 38th annual Symposium on Simulation ANSS N., & Nemoto, Y. (2007). A Stable Routing Protocol to Sup-
’05, Apr. 2005 (pp. 312–322). port ITS Services in VANET Networks. IEEE Transactions
Studer, A., Luk, M., Perrig, A, (n.d.). Efficient Mechanisms
TVT.2007.906873
to Provide Convoy Member and Vehicle Sequence Authen-
tication in VANETs. Tan, K., Zhang, Q., & Zhu, W. (2003). Shortest path routing
in partially connected ad hoc networks. In Proceedings of
SUMO Simulation of Urban Mobility. (2009). Retrieved
IEEE Globecom 2003, (Vol. 2, pp. 1038–1042).
from http://sumo.sourceforge.net/
Tariq, M. M. B., Ammar, M., & Zegura, E. (2006). Message
Sun Microsystems Inc. (1999, November). JINI Specification
ferry route design for sparse ad hoc networks with mobile
v1.0.1. Retrieved from http//:www.sun.com/JINI/
nodes. In Proceedings of the seventh ACM international
Sun Microsystems Inc. (2000). Java Media Framework API symposium on Mobile ad hoc networking and computing
Guide, 2000. Retrieved from http://java.sun.com/products/ MobiHoc ’06 (pp. 37–48).
java-media/jmf/
Thangavelul, A., & Bhuvaneswari, K. Kumar’, K., Senthil-
Sun, W., Yamaguchi, H., Yukimasa, K., & Kusumoto, S. Kumarl, K., & Sivanandam, S.N., (2007). Location Identi-
(2006). GVGrid: A QoS Routing Protocol for Vehicular Ad fication and Vehicle Tracking using VANET (VETRAC).
Hoc Networks. IEEE International Workshop on Quality In IEEE International Conference on Signal Processing,
of Service, (pp. 130-139). Communications and Networking 2007.
Sun, Y., Huang, W. L., Tang, S. M., Qiao, X., & Wang, The Network Simulator. ns2.(2009). Retrieved from http://
F. Y. (2007). Design of an OSEK/VDX and OSGi-based www.isi.edu/nsnam/ns/
embedded software platform for vehicular applications.
The zebranet wildlife tracker. (2009). Retrieved from http://
In Proceedings of the IEEE International Conference on
www.princeton.edu/ mrm/zebranet.html
Vehicular Electronic and Safety, (pp. 1-6).
The, I. E. E. E. 802.15.4 WPAN Group. (2009). Retrieved
Sun, Y., Wang, F. Y., Wang, Z. X., Qiao, X., & Wang, K.
from http://www.ieee802.org/15/pub/TG4.html
F. (2006). A scheduling algorithm for vehicular applica-
tion specific embedded operating systems. In Proceedings The, O. N. WORLD. (2005). Retrieved from http://www.
of IEEE International Conference on Systems, Man and onworld.com
Cybernetics, (pp. 2535-2540).
Third Generation Partnership Project [3GPP]. (2009).
Sven Jaap, M. B., & Wolf, L. (2005) “Evaluation of rout- Retrieved from http://www.3gpp.org
ing protocols for vehicular ad hoc networks in city traffic
Thompson, J. P. (1998). Web-based enterprise management
scenarios,” in in Proc. of the 5th International Conference
architecture. IEEE Communications Magazine, 36(3),
on Intelligent Transportation Systems Telecommunica-
80–86. doi:10.1109/35.663331
tions (ITST).
Tier. (2009). Retrieved from http://tier.cs.berkeley.edu/
Symington, S., Farrell, S., Weiss, H., & Lovell, P. (2007).
wiki/Home
Bundle security protocol specification. Internet-draft.
TIGER (Topologically Integrated GEographic Encoding
and Referencing). (2009). http://www.census.gov/geo/
www/tiger/.
394
Tomkewitsch, R. v. (1991). Dynamic route guidance and UPnP Forum. (2009). Retrieved from http://www.upnp.
interactive transport management with ALI-SCOUT. org
Journal of IEEE Transactions on Vehicular Technology,
Vahdat, V., & Becker, D. (2000). Epidemic routing for
40(1), 45–50. doi:10.1109/25.69971
partially connected ad hoc networks. Technical Report
Tonguz, O. K., Wisitpongphan, N., Parikh, J. S., Bai, F., CS-200006, Duke University.
Mudalige, P., & Sadekar, V. K. (2006). On the Broadcast
Vahid, F., & Givargis, T. (2002). Embedded system design:
Storm Problem in Ad hoc Wireless Networks. International
A unified hardware/software introduction. Hoboken, NJ:
Conference on Broadband Communications, Networks and
Wiley.
Systems, (pp. 1-11).
Wan, S., Tang, J., & Wolff, R. S. (2008). Reliable Routing
Tonguz, O., Wisitpongphan, N., Bai, F., Mudalige, P., &
for Roadside to Vehicle Communications in Rural Areas.
Sadekar, V. (2007). Broadcasting in VANET. Mobile Net-
In IEEE International Conference on Communications,
working for Vehicular Environments, (pp. 7-12).
(pp. 3017-3021).
Torge, W. (1991). Geodesy. Berlin: Walter de Gruyter.
Wang, C., Sohraby, K., Jana, R., Ji, L., & Deneshmand,
Traffic Message Channel. (n.d.). Retrieved September M. (2008). Voice communications over ZigBee networks.
11, 2009, from http://en.wikipedia.org/wiki/Traffic_Mes- IEEE Communications Magazine, 121–127. doi:10.1109/
sage_Channel. MCOM.2008.4427240
TRANSIMS. http://transims.tsasa.lanl.gov/. Wang, F. Y., Li, J. X., & Liu, X. J. (2004). OSEK/VDX and
OSGi-based Embedded Platform for Vehicular Software
Transportation Research Board. (2004). Transportation In
and its Key Techniques. Chinese High Technology Letters,
An Aging Society: A Decade of Experience. Transportation
(pp.131-136).
Research Board.
Weber, S., Cahill, V., Clarke, S., & Haahr, M. (2003). Wire-
Treiber, M., Hennecke, A., & Helbing, D. (2000) “Congested
less ad hoc network for dublin: A large-scale ad hoc network
traffic states in empirical observations and microscopic
test-bed. ERCIM News, 54. Retrieved from http://www.
simulations,” IPhys. Rev., vol. 62, no. 2.
ercim.org/publication/ErcimNews/enw54/weber.html
Tropos networks. (2009). Retrieved from http://www.
Weiss, E., Gehlen, G., Lukas, S., & Rokitansky, C. (2006).
tropos.com
MYCAREVENT- Vehicular Communication Gateway for
Tsai, H. M., Tonguz, O. K., Saraydar, C., Talty, T., Ames, M., Car Maintenance and Remote Diagnosis. In Proceedings
& Macdonald, A. (2007). ZigBee-based intra-car wireless of the IEEE International Conference on Computers and
sensor networks: A case study. IEEE Wireless Communica- Comunication, (pp. 318-323).
tions, (pp. 68-77).
WiChorus, Inc. (2009). Retrieved from http://www.musta-
Tseng, Y. C., Chen, J. J., & Cheng, Y. L. (2007). Design and faergen.com/asn_gateway.pdf
Implementation of a SIP-based Mobile and Vehicular Wire-
Wills, J., Ye, W., & Heidemann, J. (2006). Low-power
less Network with PUSH Mechanism. IEEE Transactions
acoustic modem for dense underwater sensor networks. In
Proceedings of the First ACM International Workshop on
TVT.2007.906986
UnderWater Networks (WUWNet), Los Angeles, CA, (pp.
Tseng, Y. C., Ni, S. Y., Chen, Y. S., & Sheu, J. P. (2002). 79–85). New York: ACM. Retrieved from http://www.isi.
The Broadcast Storm Problem in a Mobile Ad Hoc Net- edu/ohnh/PAPERS/Wills06a.html
work. [WINET]. ACM Wireless Networks, 8(2), 153–167.
WirelessCar website (n.d.). Retrieved September 20 2009,
doi:10.1023/A:1013763825347
from http://www.wirelesscar.com/
Unwired wireless. (2009). Retrieved from http://www.
Wisitpongphan, N., Bai, F., Mudalige, P., Sadekar, V., & Ton-
unwired.com.au/
guz, O. (2007). Routing in sparse vehicular ad hoc wireless
395
networks. IEEE Journal on Selected Areas in Communica- Hudson Valley Transportation Management Center (HVT-
tions, 25(8), 1538–1556. doi:10.1109/JSAC.2007.071005 MC) (n.d.). Retrieved from http://www.hudsonvalleytraveler.
com/Glossary.html
Wisitpongphan, N., Tonguz, O. K., Parikh, J. S., Mudalige,
P., Bai, F., & Sadekar, V. (2007). Broadcast Strom Mitiga- Open Service Gateway Initiative (OSGi). (2009). Retrieved
tion Techniques in Vehicular Ad Hoc Networks. IEEE from http://www.osgi.org
Wireless Communications, 14(6), 84–94. doi:10.1109/
WiMAX Forum. (2007). WiMAX End-to-End Network
MWC.2007.4407231
Systems Architecture Stage 3: Detailed Protocols and
Wood, L., Eddy, W. M., Ivancic, W., McKim, J., & Jackson, Procedures, Release 1.1.0.
C. (2007). Saratoga: a delay-tolerant networking conver-
WiMAX Forum. (n.d.). Wimax deployment consider for fixed
gence layer with efficient link utilization. In Proceeding
wireless Access in 2.5GHz and 3.5GHz licensed bands.
of IWSSC’07.
Xiang, W., Richardson, P., & Guo, J. (2006). Introduction
Wood, L., McKim, J., Eddy, W., Ivancic, W., & Jackson, C.
and preliminary experimental results of wireless access for
(2008). Saratoga: a convergence layer for delay-tolerant
vehicular environments (WAVE) systems. In Proceedings
networking. Internet draft.
of 3rd Annual International Conference on Mobile and
Wood, L., Peoples, C., Parr, G., Scotney, B., & Moore, A. Ubiquitous Systems, (pp.1-8).
(2007). Tcps protocol radius:the distance where timers
Xu, Q., Mak, T., Ko, J., & Sengupta, R. (2004). Vehicle-to-
prevent communication. In Proceeding of IWSSC’07.
Vehicle Safety Messaging in DSRC. In ACM International
Woodrow, B., & Thomas, A. (1998). Dingus Human Fac- Workshop on Vehicular Ad Hoc Networks (VANET), (pp.
tors in Intelligent Transportation Systems. Mahwah, NJ: 19-28).
Lawrence Erlbaum Associates.
Xu, Y. N., Kim, Y. E., Cho, K. J., Chung, J. G., & Lim, M.
World Geodetic System. (n.d.). Retrieved September 11, S. (2008). Implementation of FLexRay communication
2009, from http://en.wikipedia.org/wiki/World_Geo- controller protocol with application to a robot system. In
detic_System Proceedings of 15ht International Conference on Electron-
ics, Circuits and Systems, (pp. 994-997).
WTP1. 0 Specification (2004). Electronics and Telecom-
munications Research Institute (ETRI), Korea. Yamashita, T., Izumi, K., & Kurumatani, K. (2004). Car
navigation route information sharing for improvement of
Wu, H., Fujimoto, R., Guensler, R., & Hunter, M. (2004,
traffic efficiency. Paper presented at the Proceeding of 7th
Oct.). Mddv: a mobility-centric data dissemination al-
international IEEE conference on Intelligent Transportation
gorithm for vehicular networks. Paper presented at the
Systems, Washington, D.C., USA.
Proceedings of ACM Vehicular ad hoc networks (VANET),
Philadelphia, PA. Yang, H., Luo, H., Ye, F., Lu, S., & Zhang, L. (2004). Security
in mobile ad hoc networks: Challenges and solutions. IEEE
Wu, H., Fujimoto, R., Hunter, M., & Guensler, R. (2005).
Wireless communications, 11(1), 38-47.
An architecture study of infrastructure-based vehicular net-
works. In ACM MSWiM, Montreal, Canada, (pp.36-39). Yang, P., & Chuah, M. C. (2006). Context-aware multicast
routing scheme for disruption tolerant networks. In Pro-
Wunderlich, K. E., Kaufman, D. E., & Smith, R. L.
ceedings of ACM international workshop on PE-WASUN
(2000). Link travel time prediction for decentralized
’06 (pp. 66–73).
route guidance architectures. Journal of IEEE Transac-
tion on Intelligent Transportation Systems, 1(1), 4–14. Ye, Q., Cheng, L., Chuah, M. C., & Davison, B. D. (2006).
doi:10.1109/6979.869017 Os-multicast: On-demand situation-aware multicasting
in disruption tolerant networks. In . Proceedings of IEEE
WWWOFFLE. (2009). Retrieved from http://www.gedan-
VTC, 1, 96–100.
ken.demon.co.uk/wwwoffle/
396
Yokomori, I., & Suzuki, Y. (2003). Pressure sensors. In J. Zhao, J., & Cao, G. (2006). VADD: Vehicle-Assisted Data
Marek, H.P. Trah, Y. Suzuki,& I. Yokomori, (Ed.), Sensors Delivery in Vehicular Ad Hoc Networks. IEEE Computer
application Vol. 4 - Sensors for automotive technology (pp. Communications, (pp. 1-12).
Zhao, W., Ammar, M., & Zegura, E. (2004). A message
Yoon, H., Kim, J., Tan, F., & Hsieh, R. (2008). On-demand ferrying approach for data delivery in sparse mobile ad
Video Streaming in Mobile Opportunistic Networks. In hoc networks. In Proceedings of the 5th ACM international
IEEE International Conference on Pervasive Computing symposium on Mobile ad hoc networking and computing
and Communications (PerCom), (pp. 80–89). MobiHoc ’04 (pp. 187–198).
Yousefi, S., Mousavi, M. S., & Fathy, M. (2006). Vehicular Zhao, W., Ammar, M., & Zegura, E. (2005). Controlling
Ad Hoc Networks (VANETs): Challenges and Perspectives. the mobility of multiple data transport ferries in delay-
International conference on ITS Telecommunications, (pp. tolerant network. In . Proceedings of IEEE INFOCOM,
761-766). 2, 1407–1418.
Yu, F., & Biswas, S. (2007). Self-Configuring TDMA Zhao, W., Ammar, M., & Zegura, E. (2005). Multicasting
Protocols for Enhancing Vehicle Safety with DSRC Based in delay tolerant networks: Semantic models and routing
Vehicle-to-Vehicle Communications. IEEE Journal on algorithms. In Proceeding of Sigcomm Workshop in DTN,
Selected Areas in Communications, 25(8), 1526–1537. (pp. 268–275).
doi:10.1109/JSAC.2007.071004
Zhao, Y. (1997). Vehicle Location and Navigation Systems.
Zeng, X., Bagrodia, R., & Gerla, M. (1998). Glomosim: A Boston: Artech House.
library for the parallel simulation of large-scale wireless
Zheng, N. N., Tang, S., Cheng, H., Li, Q., Lai, G., & Wang,
networks. In The 12th Workshop on Parallel and distrib-
F. W. (2004). Toward Intelligent Driver-Assistance and
uted Simulation. PADS’98, Banff, Alberta, Canada (pp.
Safety Warning Systems. IEEE Intelligent Systems, 19(2),
154–161).
8–11. doi:10.1109/MIS.2004.1274904
Zhan, F. B. (1997). Three fastest shortest path algorithms
Zheng, Q., Hong, X., & Ray, S. (2004). Recent advances in
on real road network: data structure and procedures.
mobility modeling for mobile ad hoc network research. In
Journal of geographic information and decision analysis,
Proceedings of the 42nd ACM annual Southeast regional
1(1), 69-82.
conference (ACM-SE), (70-75). New York: ACM Press.
Zhang, D., Xiao, H. W., & Hackbarth, K. (2004). OSGi
Zhou, Y., Wang, X., & Zhou, M. (2006). The Research and
based service infrastructure for context aware automotive
Realization for Passenger Car CAN Bus. In Proceedings
telematics. In . Proceedings of IEEE International Confer-
of the IEEE International Forum on Strategic Technology,
ence on Vehicular Technology, 5, 2957–2961.
(pp. 244-247).
Zhang, Y., Zhao, J., & Cao, G. (2007). On Scheduling
Zigbee Alliance. (2006). ZigBee specification: ZigBee
Vehicle-Roadside Data Access. In Proceedings of the ACM
Document 053474r13.
International Workshop on Vehicular Inter-Networking.
Zimmermann, H.-M., Gruber, I., & Roman, C. (2005) “A
Zhang, Z., & Fei, Z. (2007). Route design for multiple
voronoi-based mobility model for urban environments,” in
ferries in delay tolerant networks. In Proceedings of IEEE
in Proc. Of the European Wireless 2005 (EW05) .
Wireless Communications and Networking Conference
(pp. 3460–3465). Zonoozi, M. M., & Dassanayake, P. (1997). User mobility
modeling and characterization of mobility patterns . IEEE
Journal on Selected Areas in Communications, 15(7),
1239–1252. doi:10.1109/49.622908
397
398
About the Contributors
Chung-Ming Huang received the B.S. degree in electrical engineering from National Taiwan Uni-
versity on 1984/6, and the M.S. and Ph.D. degrees in computer and information science from The Ohio
State University on 1988/12 and 1991/6. Currently, he is a Distinguished Professor of Dept. of Computer
Science and Information Engineering, National Cheng Kung University, Taiwan, R.O.C. He also serves as
(i) Director of the Promotion Center for the Telematics Consortium (PCTC), Ministry of Education (MOE),
Taiwan, R.O.C. and (ii) Principal Project Reviewer of Industrial Development Bureau and Department of
Industrial Technology, Ministry of Economic Affairs (MOEA), Taiwan, R.O.C. He has published more than
200 referred journal and conference papers in wireless and mobile communication protocols, interactive
multimedia systems, audio and video streaming and formal modeling of communication protocols. His
research interests include wireless and mobile network protocol design and analysis, media processing
and streaming, web technologies, and network applications and services.
Yuh-Shyan Chen received the B.S. degree in Computer Science from Tamkang University, Taiwan,
R. O. C., in June 1988 and the M.S. and Ph.D. degrees in Computer Science and Information Engineer-
ing from the National Central University, Taiwan, R. O. C., in June 1991 and January 1996, respectively.
He joined the faculty of Department of Computer Science and Information Engineering at Chung-Hua
University, Taiwan, R. O. C., as an associate professor in February 1996. He joined the Department of
Statistic, National Taipei University in August 2000, and joined the Department of Computer Science and
Information Engineering, National Chung Cheng University in August 2002. Since 2006, he has been a
Professor at the Department of Computer Science and Information Engineering, National Taipei University,
Taiwan. Prof. Chen is now serving as chair of Institute of Communication Engineering, National Taipei
University, Taiwan, ROC, and Vice Chair of Task Force on “Telecommunications” of Intelligent Systems
Applications Technical Committee, IEEE Computational Intelligence Society from 2007. Prof. Chen
served as Editor-in-Chief of International Journal of Ad Hoc and Ubiquitous Computing (SCIE), Editorial
Board of Telecommunication System Journal (SCIE), EURASIP Journal on Wireless Communications
and Networking (SCIE), and Mobile Information Systems (SCIE). He served as Guest Editor of ACM/
Springer Mobile Networks and Applications (MONET), Telecommunication Systems, Wireless Commu-
nications and Mobile Computing, EURASIP Journal on Wireless Communications and Networking, The
Computer Journal, Wireless Personal Communications, International Journal of Communication Systems,
and IET Communications. His paper wins the 2001 IEEE 15th ICOIN-15 Best Paper Award. Prof. Chen
was a recipient of the 2005 Young Scholar Research Award, National Chung Cheng University, R.O.C..
His recent research topics include wireless communications, mobile computing, and next-generation
personal communication system. Dr. Chen is a senior member of the IEEE Communication Society and
Phi Tau Phi Society.
***
Ben-Jye Chang received his M.S. degree in computer engineering from University of Massachu-
setts, Lowell, USA, in 1992 and the Ph.D. degree in computer science and information engineering from
National Chung-Cheng University, Taiwan, in 2001. He is currently a full professor in the Institute of
Computer Science and Information Engineering, National Yunlin University of Science and Technol-
ogy, Taiwan. He was the Chair of the Department of Computer Science and Information Engineering
of the Chaoyang University of Technology, Taiwan, in 2008, where he joined the department in 2002.
His research interests include seamless vertical handoff in heterogeneous wireless networks, cross
layer-based mechanism, resource management and analysis in WiMAX, adaptive-based wireless TCP
congestion control, mobile computing in wireless sensor networks, GPS-based adaptive routing in
VANET, time and frequency 2-D spreading in next generation cellular communications, and QoS-based
wireless networking.
Chih-Yung Chang is a Full Professor of Department of Computer Science and Information En-
gineering at Tamkang University, Taiwan. Dr Chang served as an Associate Guest Editor of Journal
of Internet Technology (JIT, 2004 and 2008), Journal of Mobile Multimedia (JMM, 2005), Telecom-
munication Systems (TS, 2009), Journal of Information Science and Engineering (JISE, 2009) and a
member of Editorial Board of International Journal of Ad Hoc and Ubiquitous Computing (IJAHUC,
2009), Journal of Information Technology and Applications (JITA, 2008-2009), and Tamsui Oxford
Journal of Mathematical Sciences (2001-2009). Dr. Chang is a member of the IEEE Computer Society,
Communication Society and IEICE society. His current research interests include vehicular ad hoc
networks, wireless sensor networks, WiMax broadband networks, Bluetooth radio networks, Ad Hoc
wireless networks, and mobile computing.
Jen-Chun Chang is a Full Professor of Department of Computer Science and Information Engineer-
ing, National Taipei University, Taipei, Taiwan. Professor Chang received the B.S. and M.S. degrees
in Computer Science and Information Engineering from National Taiwan University, Taipei, Taiwan,
in 1989 and 1991, respectively. His Ph.D. degree in Computer Science and Information Engineering
was received from National Chiao Tung University, Hsinchu, Taiwan, in 2000. He has co-authored two
books (in English versions): Systems and Bayesian Reliability (published by World Scientific, 2002),
Handbook of Reliability Engineering (published by Springer, 2003). Recently, most of his research results
are published in IEEE Transactions on Information Theory and Designs, Codes, and Cryptography. His
research interests include coding theory, cryptography, reliability, and algorithms.
Teng-Wen Chang received the M.S degree in computer science and Computer Science and Informa-
tion Engineering from Nation Dong Hwa University, Taiwan, Republic of China, in July 2009. In current,
he studies at Department of Electrical Engineering from Nation Taiwan University, Taiwan, Republic
of China for PhD degree. His research interests include digital home /telematics network, quantum key
distribution, cloud computing, and NetFPGA enhanced network.
Yao-Chung Chang received B.S. degree (1996) of Computer Science and Information Engineering
from Tamkang University, Taipei, Taiwan, M.S. degree (1998) and Ph.D. degree (2006) of Computer
Science and Information Engineering from National Dong Hwa University, Hualien, Taiwan. He is an
Assistant Professor now in the Department of Computer Science and Information Engineering, National
Taitung University, Taitung, Taiwan. Also, he serves as the Director of Computer Center, National
399
Taitung University. His main research focuses on the RFID/EPC Network, Sensor Network, Network
Mobility, Network Planning & Management and Transition of IPv4/IPv6.
Yi-Shiang Chang received the B.S. degree in Department of Information Management from Shih
Hsin University, Taiwan, in June 2004 and the M.S. degree in Department of Computer Science and
Information Education from National University of Tainan, Taiwan, in July 2006.Since September 2007,
he has been working towards the Ph.D. degree and currently is a doctoral candidate in the Department of
Computer Science and Information Engineering, National Cheng Kung University, Taiwan. His research
interests include mobile computing, wireless networking, and mobile learning.
Ching-Wen Chen received the M.S. degree in the Department of Computer Science from National
Tsing Hwa University, Taiwan, 1995. He obtained his Ph.D. in Computer Science and Information En-
gineering from the National Chiao-Tung University, Taiwan, 2002. He was an Assistant Professor at
the Chaoyang University of Technology (2005–2007), Taiwan. Currently, he is an Associate Professor
in the Department of Information Engineering and Computer Science, Feng Chia University, Taiwan.
His research interests include computer architecture, parallel processing, embedded system, mobile
computing, and wireless sensor network.
Jiann-Liang Chen Professor Chen received the Ph.D. degree in Electrical Engineering from National
Taiwan University, Taipei, Taiwan in 1989. Since August 1997, he has been with the Department of
Computer Science and Information Engineering of National Dong Hwa University, where he is a profes-
sor and Vice Dean of Science and Engineering College. Prof. Chen joins the Department of Electrical
Engineering, National Taiwan University of Science and Technology, as a full professor and Vice Dean
of the College of Electrical and Computer Engineering now. He has published more than 100 papers in
journals and conferences, and also holds several patents. His current research interests are directed at
mobile computing, digital home network, telematics applications, and RFID middleware design.
Ming-Chiao Chen received the Ph.D. degree in Computer Science and Information Engineering
from National Dong Hwa University, Hualien, Taiwan, in 2008. He joins the Department of Computer
Science and Information Engineering, National Taitung University, Taitung, Taiwan, since 2008 and is
now an assistant professor. His current research interests are in the area of wireless networks, sensor
networks and cross-layer design.
Min-Xiou Chen is an Assistant Professor of Department of Computer Science and Information

Engineering, National Dong Hwa University, Hualien, Taiwan, ROC from 2008. Prof. Chen is received
the B.S. degree in Computer and Information Science from Tung Hai University, Tai-Chung, Taiwan,
ROC in 1996, and M.S. and Ph.D. degrees in Computer Science and Information Engineering from
National Chung Cheng University, Chia-Yi, Taiwan, ROC in 1998 and 2005, respectively. In August
2005, Prof. Chen joined the facutly of Department of Computer Science and Information Engineering,
Chung Hua University, as an Assistant Professor. Prof. Chen also is an IEEE member. His recent re-
search topics include wireless communications, wireless sensor network, VANET, and next-generation
personal communication system.
400
Tzung-Shi Chen received the Ph.D. degree in Computer Science and Information Engineering from
National Central University, Taiwan, in June 1994. Currently, he is a Full Professor at the Department
of Computer Science and Information Engineering, National University of Tainan, Tainan, Taiwan. He
has served as a Guest Editor of Journal of Internet Technology and International Journal of Ad Hoc
and Ubiquitous Computing. He has also served as PC members on many international conferences. His
current research interests include mobile computing and wireless networks, ubiquitous and pervasive
computing, and mobile learning. Dr. Chen is a member of the IEEE Computer Society and IEEE Com-
munications Society.
Wei-Kuo Chiang was born in Taiwan, 1967. He received B.S., M.S., and Ph.D. degrees in Computer
Science and Information Engineering from National Chiao Tung University (NCTU), Taiwan, R.O.C.,
in 1989, 1991 and 1996, respectively. In February 2004, he joined the Department of Computer Science
and Information Engineering, National Chung Cheng University, Chiayi, Taiwan, as an Assistant Pro-
fessor. Before that, he was a Section Manager of the Internet Telecommunications Department at the
Computer and Communications Research Laboratories, Industrial Technology Research Institute (CCL/
ITRI), Taiwan. Dr. Chiang holds three patents with six pending. His recent research interests include
mobility management in IP multimedia subsystems, service technologies in next generation networks,
mobile computing and wireless networks.
Cheng Hsuan Cho received the B.S. degree in information Management from the Chang Cung
University, Taiwan, Republic of China, in June 2007 and the M.S. degree in communication engineering
from National Chung Cheng University, Taiwan, Republic of China, in June 2009. His research interests
include wireless network and Vehicular ad-hoc network.
Chyi-Ren Dow was born in 1962. He received the B.S. and M.S. degrees in information engineering
from National Chiao Tung University, Taiwan, in 1984 and 1988, respectively, and the M.S. and Ph.D.
degrees in computer science from the University of Pittsburgh, USA, in 1992 and 1994, respectively.
Currently, he is a Professor in the Department of Information Engineering and Computer Science, Feng
Chia University, Taiwan. He served as Guest Editor of Journal of Pervasive Computing and Communi-
cations (JPCC) and International Journal of Computer Processing of Oriental Languages. His research
interests include mobile computing, ad-hoc wireless networks, telematics, software agents, and learning
technologies.
Chun-I Fan was born in Tainan, Taiwan. He received his M.S. degree in computer science and in-
formation engineering from National Chiao Tung University, Taiwan, in 1993, and the Ph.D. degree in
electrical engineering at National Taiwan University in 1998. From 1999 to 2003, he was an associate
researcher and project leader of Telecommunication Laboratories, Chunghwa Telecom Co., Ltd, Tai-
wan. In 2003, he joined the faculty of the department of computer science and engineering, National
Sun Yat-sen University, Kaohsiung, Taiwan, and is an associate professor now. He won the Dragon
Thesis Award from Acer Foundation and Best Thesis Award from Institute of Information & Comput-
ing Machinery in 1999, Best Student Paper Awards in National Conference on Information Security
1998 and 2007. His current research interests include information security, cryptographic protocols,
wireless security, and electronic commerce, and he has published over 80 papers in journals, books,
and conference proceedings.
401
Jenq-Muh Hsu received his M.S. and Ph.D. degrees in computer science and information engi-
neering from National Cheng-Kung University, Taiwan, in 1993 and 1998, respectively. He joined the
Department of Computer Science and Information Engineering, National Chia-Yi University, Chia-Yi
City, Taiwan, in 2009, where he is now an Assistant Professor. His research interests include mobile
computing application services, ubiquitous computing, and e-learning.
Ruei-Hau Hsu was born in Kaohsiung, Taiwan on July 11, 1979. He received his B.S. and M.S.
degrees in Computer Science and Information engineering from Tunghau University, Taiwan, in 2002
and 2004, respectively. He is toward to his Ph.D. degree in Computer Science and Information Engi-
neering at National Sun Yat-sen University from 2005 until now. His current research interests include
information security and cryptographic primitives (i.e., wireless authentication protocols and signature
schemes).
Kun-Chan Lan is an Assistant Professor of Department of Computer Science and Information

Engineering, National Cheng Kung University, Taiwan, Republic of China. He received his B.S degree
in industrial management science from National Cheng Kung University, Taiwan in 1990. He received
his MS degree in Computer Science from State University of New York at Stony Brook, USA in 1997.
and his Ph.D degree in Computer Science from University of Southern California, USA in 2004. He was
listed in the Marquis Who\’s Who in the World, 2009. He participated in many International conferences
and actively served the research community, such as Technical Program Committee of the Vehicular
Electronics & Telematics track, Technical Program Committee of the IEEE Vehicular Networks &
Applications Workshop, Program Committee (PC) for the First International Workshop on Vehicular
Communication Technologies (VehiCom 2009), etc. He also organized several international workshop,
including the 1st and the 2nd IEEE International Workshop on Opportunistic Networking (WON), the
1st and the 2nd IEEE International Workshop on Vehicular Networking (VON). Professor Lan’s research
interests include Realistic Network Simulation, Vehicular Network, Wireless Sensor Network.
Da-Jie Lin was born in 1968. He received the B.S. degree in civil engineering from National Taiwan
University, Taiwan, in 1990, the M.S. degree in civil engineering from Cornell University, USA, in 1996,
and the Ph.D. degree in civil engineering from U.C. Berkeley, USA, 2001. Currently, he is an assistant
professor in the Department of Transportation Technology and Management, Feng Chia University, Tai-
wan. His research interests include intelligent transportation systems, traffic control, and algorithms.
Cheng-Min Lin received the B.S. and M.S. degrees in electronic engineering from National Taiwan
University of Science and Technology, Taipei, Taiwan, in 1989 and 1991, respectively, and the Ph.D. de-
gree in Department of Information Engineering and Computer Science, Feng-Chia University, Taichung,
Taiwan. Currently, he is an Associate Professor in the Department of Computer and Communication
Engineering, Graduate Institute of Electrical Engineering and Computer Science, Nan Kai University
of Technology, R.O.C.. His recent research interests include embedded systems, mobile computing,
distributed systems, wireless sensor networks and telematics. Dr. Lin is a member of the IEICE and a
member of the IEEE Computer Society, respectively.
402
Jyh-Horng Lin was born in 1975. He received the B.S., M.S., and Ph.D. degrees in information
engineering from Feng Chia University, Taiwan, in 1998, 2000, and 2005, respectively. Currently, he
is an Assistant Professor in the Department of Electrical Engineering, Nan Kai University of Technol-
ogy, Taiwan. His research interests include mobile computing, ad-hoc wireless networks, distributed
systems, and computer algorithms.
Yun-Wei Lin received the B.S. degree in Computer and Information Science from the Aletheia
University, Taiwan, R.O.C., in June 2003 and the M.S. degree in Computer Science and Information En-
gineering from National Chung Cheng University, Taiwan, R.O.C., in July 2005. He is currently working
toward the Ph.D. degree in computer science and information engineering, at the National Chung-Cheng
University, Taiwan, R.O.C. He has received the CSIE Ph.D candidate from the National Chung Cheng
University, 2009. His research interests include mobile ad hoc network, and wireless sensor network.
Tzong-Jye Liu received the PhD degree in 1999 from the Department of Computer Science, National
Tsing Hua University, Taiwan. After he got his PhD degree, he worked several years in the computer
industry in Taiwan. He was an Assistant Professor at the Feng Chia University (2004–2008), Taiwan.
Dr. Liu is currently an Associate Professor at the Department of Information Engineering and Computer
Science, Feng Chia University, Taiwan. His research interests include operating systems, distributed
computing and network security
Jen-Yi Pan received the B.S. and Ph.D. degrees in computer science from National Tsing-Hua Uni-
versity, Hsinchu, Taiwan, R.O.C., in 1995 and 2002, respectively. He is currently an assistant professor
with the Department of Communications Engineering and with the Center for Telecommunication
Research, National Chung Cheng University, Chaiyi, Taiwan, R.O.C. His research interests include
performance evaluation of medium access control and Internet telephony. Dr. Pan is a member of ACM,
IEEE, and IEICE.
An-Nie Ren received B.S. and M.S. degrees in Computer Science and Information Engineering
from National Chung Cheng University, Taiwan, R.O.C., in June 2007 and July 2009, respectively. Her
research topics include SIP mobility in 3GPP IP Multimedia Subsystem, SIP-based network mobility,
and interworking between WLAN and 3G cellular networks.
Hua-Wen Tsai received the B.S. degree in Information Management from Chang Jung Christian
University, Taiwan, in June 1998 and the M.B.A. degree in Business and Operations Management from
Chang Jung Christian University, Taiwan, in June 2001, and the Ph.D. degree in Computer Science and
Information Engineering, National Cheng Kung University, Taiwan, in April 2008. He is joined the
faculty of the Department of Computer Information and Network Engineering, Lunghwa University
of Science and Technology, Taoyuan, Taiwan, as an Assistant Professor in August 2008. His current
research interests include wireless communications, ad hoc networks, and sensor networks.
Chung-Ping Young received the B.S. degree in electronic engineering from Chung-Yuan Christian
University, Taiwan, in 1985, and the M.S. and Ph.D. degrees in electrical engineering from the University
of Missouri, Columbia, in 1994 and 1997, respectively.From 1994 to 1997, he was a Research Assistant
403
at the University of Missouri, where he worked on power measurement. From 1998 to 2003, he was with
Wistron InfoComm and Phoenix Technologies as a Senior Engineer and Principal Engineer. In 2003,
he joined the Department of Computer Science and Information Engineering, National Cheng-Kung
University, Taiwan, as an Assistant Professor. His fields of interest include telematics, wireless sensor
network, multi-core embedded software, and biomedical instrumentation.
404
405
Index
A Automatic Vehicle Classification (AVC) 34

automatic vehicle identification (AVI) 34, 38
AAA framework 148
automatic vehicle locations (AVL) 34, 36, 38
accelerometer 69
Automatic Vehicle Monitoring (AVM) 36, 38
access services network (ASN) 8
ACC system 62 B
adaptive cruise control (ACC) 58, 61
Adaptive traffic control systems 331 backbone creation 203
ad hoc domain 372 backbone member 203
Ad-hoc On-demand Distance Vector (AODV) Backbone Member (BM) 202
80 Back-to-Back UA (B2BUA) 178
Advanced Public Transportation Systems base station (BS) 8
(APTS) 35 beacon-enabled network 79
American Electronic Sign Co (AES) 54 Bellman-Ford algorithm 135
analog-to-digital converter (ADC) 68 biometric system 72
Android Market 294 blind-spot warning 60
Android/OSGi applications 305, 306, 307 block diagram 64, 65, 66
Android platform 290, 291, 292, 294, 296, Bluetooth 167, 181
299, 300, 301, 304, 307, 308 broadcast protocols 208, 226
Android-powered infotainment operating sys-
C
tem 292
Android software 292 C2F protocols 48, 49
antilock braking system (ABS) 58, 60 C2X-applications 248
application management 15, 21 CAN (Controller Area Network) 1, 2
application programming interface (APIs) 290 Car alarm system 124
application server (AS) 164 CAR protocol 211, 212, 213, 214
Application Support Sublayer (APS) 77 Center-to-Center (C2C) 32
APS data entity (APSDE) 77 central processing unit (CPU) 232
APS management entity (APSME) 77 changeable message signs (CMS) 33, 34, 35
APTS system 36 CIM Object Manager (CIMOM) 315
architecture framework 144 CIM schema 314, 315
ASN-anchored mobility 149, 150 CIM specification 314
Aspect-Oriented Programming (AOP) 298 Cologne-based project 297
asynchronous communication 258 Common Gateway Interface (CGI) 24
Automatic Cargo Identification (ACI) 38 common information model object manager
automatic number plate recognition (ANPR) (CIMOM) 322
33
Index
common information module (CIM) 320 Distance-Driven Probabilistic Diffusion

communication protocols 160, 161, 162, (DDPD) 199
167, 171, 173, 174, 175 D-MinCost algorithm 217
communications media 51 Document Type Definition (DTD) 315
complex network protocols 230 DSRC protocol layer 184
components-oriented 295 DTN routing algorithms 256
computer vision system 128, 129, 130 dynamic aspect 119
computing environment 295 dynObjIndex 281, 282, 285
Confidentiality block (CB) 264
connectivity services network (CSN) 8 E
contention access period (CAP) 78 Earliest Delivery with All Queues (EDAQ) 258
context-aware multicast routing (CAMR) 264 Earliest Delivery with Local Queuing (EDLQ)
context-aware multicast routing (CAMR) 258
scheme 264 effective path length (EPL) 259
Context-Aware Routing (CAR) protocol 258 electronic control unit (ECU) 4, 6, 58, 61, 296
context information 17, 18, 28 Electronic Emergency Barke Light (EEBL)
control channel (CCH) 91, 95 372
Controlled Vehicular Internet Access (CVIA) Electronic Fare Payment (EFP) 36
185 electronic stability program (ESP) 58, 66
controller area network (CAN) 2, 5, 20, 75 Electronic Toll Collection (ETC) 34, 38
CVO technologies 38 e-map 119
cyclic redundancy checksum (CRC) 4 embezzlement 124
emergency-vehicle-approach 225
D
encounter-based multicast routing (EBMR)
data encryption mechanism 323 264
data packets 82, 83 end-to-end communicate 100
data transmission 23 end-to-end paths 232
data transmissions 206, 261 environmental information 65
Dedicated short range communications (DSRC) Evaluate state 193
231 Extensible Authentication Protocol (EAP) 148
Delay-bounded routing protocol 209
Delay/disruption tolerant networks (DTN) 252 F
delay model 211 Federal Communications Commission (FCC)
Delay Tolerant Network (DTN) 252, 254 231
Denial-of-service (DoS) 377 Felix framework 304
Department of Defense (DoD) 109 ferry relaying algorithm (FRA) 259
diagnosis system 310, 319, 327 Fixed Mobile Convergence (FMC) 163
Diamond Digital router 339 FlexRay protocol 2, 5
digital data 61, 68 full function device (FFD) 76
digital library 133, 135 Function-Driven Probabilistic Diffusion
digital map database 131, 133, 134, 137, 138 (FDPD) 200
digital map library 133, 135
digital multimedia renderer (DMR) 179 G
digital multimedia server (DMS) 179
GBR (GroupBased Routing) 263
Digital Right Management (DRM) 170
General Packet Radio Service (GPRS) 10
Direct Sequence Spread Spectrum (DSSS) 77
406
Index
generated-map-based model 350 IMS network 177, 178, 179, 180

Generation Telematics Protocol 28, 30 Information and Communication Technology
geocast protocol 206, 217, 218, 219 (ICT) 15
Geographic Information Systems (GIS) information module 320, 321, 322, 323
34, 35, 39 integrated circuit (IC) 67
geographic routing protocol 365 intelligent transportation system (ITS) 32, 33,
Global Positioning Systems (GPS) 15, 35, 247 206, 246
Global Satellite Positioning System (GPS) Inter-agent communication 18
108, 109 inter-contact time 255
global system for telematics (GST) 295 inter-departure time 359
Google Earth 356, 358, 359 International Organization of Standards (ISO)
GPS receiver 109, 110, 111, 113, 115, 116 49
gross domestic product (GDP) 369 Internet Protocol Based (IP-BASED) 318
ground-based destinations 236 Internet service provider (ISP) 320
Group-Based Routing (GBR) 263 inter-node connections 243
inter-vehicle communication (IVC) 210, 214,
H 219, 230, 370
hardware description language (HDL) 67 inter-vehicle communications system 225
Health management server (HMS) 23 intra-ASN mobility 149
Heterogeneous Vehicular Network Diagnosis in-vehicle network
System 310 1, 2, 5, 6, 58, 59, 61, 62, 68
HIGA functional component 181 in-vehicle network architecture 312
Highway Advisory Radio (HAR) 35, 39 in-vehicle networks 1, 12
Home Gateway Initiative (HGI) 160 In-Vehicle Telematics Systems (IVIS) 18
Home IMS Gateway (HIGA) 160, 177 IP-based home network 317
Home Subscriber Server (HSS) 164 IP-based multimedia application 160
home/vehicular network IP-based multimedia services 160
160, 161, 165, 170, 175, 181 IP-based network technology 161
hop-by-hop manner 203 IP Multimedia Service Control (ISC) 164
HTTP (Hyper Text Transmission Protocol) 162 IP Multimedia Subsystem (IMS)
HTTP-over-DTN 266 160, 161, 181
HTTP-over-TCP 266 IP service model 144
HTTP standard protocol 170
J
Hudson Valley Transportation Management
Center (HVTMC) 236, 249 Java Archive (JAR) 295
human-computer interaction 23 JINI location 18
human-machine interface JINI middleware 17, 18
131, 133, 136, 137, 138, 291
Human-to-human communications 47 K
Hybrid Simulation Mode 365 knowledge-based mechanisms 256, 268
hybrid vehicular communication (HVC) 230 K virtual machine (KVM) 296
hydraulic modulator 65
L
I
LAN control module 321, 322, 323
IC technology 59, 64 LAN management module
IEEE 1609 90, 91, 92, 93, 94, 95, 97, 98, 320, 321, 322, 323, 324, 325
100, 103, 104
407
Index
LAN module 320, 321, 322, 323, 324, 325 mobile vehicular network 62
LDAP (Lightweight Directory Access Protocol) mobility model 348, 349, 352, 353, 355,
166 356, 357, 360, 361, 363, 364, 365,
Licklider Transmission Protocol (LTP) 261 366, 367, 368
LIN (Local Interconnect Network) 1, 2, 312 MOST (Media Oriented Systems Transport)
Linux-based Ethernet router 339, 344 1, 2
Linux kernel 291, 292, 294, 299 MRT environment 158
local interconnect network (LIN) 2, 5, 20 multicast packets 207, 217, 226
Location centric 243 multicast tree 263
LR-WPAN 74, 75, 76 multi-hop 370, 371, 372, 375, 377
multi-hop communication 185, 207
M multi-hop forwarding
machine interaction 72 206, 209, 210, 211, 216, 217
MAC protocols multi-hop manner 183, 185
183, 184, 185, 195, 204, 253 multi-hop transmission 209
macro-scale levels 356 multi-hop wireless transmission 207
MAC sublayer management entity (MLME) multimedia contents 71
91, 92 multimedia data 176, 177, 180, 181
management information base (MIB) 95, 271 multimedia data sharing 177, 181
management object format (MOF) multimedia file 169, 170, 180
322, 327, 329 multimedia storage 161
MANET simulation 349 multiple routes (MURA) 259
map-based Activities 307
N
map-matching algorithm 119
map-matching function 134, 135, 137 navigation routing algorithm 119
map-matching module 133 navigation system 119, 125, 126, 127, 128
map-matching process 131, 133, 134, 137 Network Access Provider (NAP) 144
media oriented systems transport (MOST) 5 network architecture 1, 2, 3, 7, 8, 9
media streaming 161 Network Attached Storage (NAS) 179
medium access control (MAC) 230 network devices 161, 171
mesh node 339, 341, 342, 345 network environment 160, 167, 174
metropolitan-area wireless mesh networks network fragmentation 206, 207, 208, 214,
(MAWMN) 332 217, 218, 219, 221, 222, 225
metropolitan-scale WMN 331 network framework 163, 170, 171, 181
micro-controller unit (MCU) 11 network-layer packet 377
microelectrical mechanical system (MEMS) 69 Network Management Framework 271
mobile ad hoc networks (MANETs) 206, 207, Network-Management Systems (NMSs) 317
229, 230, 369, 371 network monitoring 7
mobile agents 25, 26 network on wheels (NoW) 371
mobile device 161, 170, 179, 180 Network protocol conversion 321, 322
mobile navigation assistant (MNA) 137 network protocol conversion module
mobile network node 7 321, 322, 324
mobile networks 2 network reference model 144, 145, 146
mobile router 24 Network Service Provider (NSP) 144
mobile station (MS) 8 network simulator 349, 354
Mobile telematics computing environment 24 Network Surveillance 288
408
Index
network topology 253, 258 Point-to-Point Protocol (PPP) 41, 52

Next Generation Network (NGN) 162, 163 position-based fashion 210
Node centric 243 Post Crash Notification (PCN) 372
Node-Initiated MF (NIMF) 259 Practical Expected Delays (PEDs) 262
non-beacon-enabled network 79 prepaid server (PPS) 158
non-real-time applications 331, 335 probabilistic-based MAC protocol 185
nonrecurring engineering (NRE) 67 Probe Surveillance 288
non-stationary solutions 242 programmable logic device (PLD) 67
Normal Vehicle (NV) 202 Protocol data unit (PDU) 279
NTCIP-based system 49 protocol stack 45, 48, 52
NTCIP-compatible traffic 41 prototype model 6
NTCIP protocols 42, 44, 49, 53, 54 push-to-talk (PTT) 84
PXE network 345
O
Q
object-oriented schema 314
on-board computers 1, 9 QoS-based accounting 157, 158
on-board computers (OBC) 298
onboard units (OBU) 38, 371 R
one-hop delay 196 radio-frequency ID (RFID 72
one-hop transmission 195 radio frequency (RF) 69
One Zero Flooding (OZF) 199 RADIUS Server 158
Open Service Gateway Initiative (OSGi) random waypoint (RWP) 232, 349
160, 181, real-time map 365
Open Service Gateway Initiative Service Plat- real-time road information selections 243
form (OSGi SP) 160 Received Channel Power Indicator (RCPI) 95
operating system (OS) 6, 67 reduced function device (RFD) 76
operation environment 63 regional computers (RC) 332
Opportunistic forwarding 243 Regional Traffic Control 288
Original Equipment Manufacture (OEM) plat- Remote Access Client (RAC) 178
forms 290 Remote Access Discovery Agent (RADA) 178
orthogonal frequency division multiplex Remote Access Server (RAS) 178
(OFDM) 231 Remote Access Transport Agent (RATA) 178
OSEK Implementation Language (OIL) 21 remote alarm system 123
OSEK/VDX operating system 2 remote transmission request (RTR) 4
OSGi framework 295, 299, 300, 301, 302, Road Feature Notification (RFN) 372
303, 305, 307, 313, 314, 325 Road Hazard Condition Notification (RHCN)
OSGi platform 161, 165, 171, 174, 310, 372
311, 312, 325, 326 Road-Look-Ahead (RLA) 17
out-vehicle network 1, 2, 3 roadside-to-vehicle communication systems
out-vehicle network architecture 1, 2, 3 (RVC) 230
OZF 199 roadside to vehicle (R2V) 213
Road Side Unit (RSU) 183, 371, 372
P
ROMSGP protocol 213
peer-to-peer topology 79 route discovery table (RDT) 82
Personal Digital Assistant (PDA) 372 route guidance process
Physical Layer (PHY) 184, 204 131, 133, 135, 136, 137
409
Index
route planning process Sub-network 321

131, 133, 134, 135, 136, 137 sub-network agent 320, 321, 323, 325
routing protocols 206, 207, 208, 209, 210, sub-network information 320, 321
218, 220, 222, 223, 226 sub-network instruction 322
routing table (RT) 82 Sydney Coordinated Adaptive Traffic System
RREP packets 82, 215 (SCATS) 333
RTA wireless network 334 system architecture 325
system-on-chip (SoC) 67
S system platform 131, 137
Sami Network Connectivity (SNC) 254
T
satellite technology 125
SDP parameters 179 TCP-based convergence layer (TCPCL) 262
service-based architecture 18 TCPCL connection 262
service channel (SCH) 95 TCP protocol 95
service environments 23 TDMA-based MAC protocols 183, 185, 204
service-integrated platform 160, 161 TDMA system 185
service-oriented architecture (SOA) 295 Telecommunications Management Network
ServiceReference array 304 (TMN) 311
shortest expected path routing (SEPR) 259 telematics protocol 15, 28
Short-range radars (SRR) 69 telematics service 2
Simple Network Management Protocol telematics system 17
(SNMP) 42, 317, 318 telemetric applications 15
Simple Transportation Management Framework TIGER maps 359, 364
(STMF) 271 topology 206, 207, 208, 210, 217, 219,
Simple Transportation Management Protocol 349, 357, 361, 363, 366
(STMP) 42, 271 trace-based 349
simulation model 352 traction control system (TCS) 58, 61, 65
SIP Proxy Server 171 traffic control system 365
SIP Service Bundle 171, 172, 173, 174, 175 traffic information center (TIC) 137
sleep mode 74, 78, 79, 87 traffic management centre (TMC) 333
Slow Vehicle Alert (SVA) 372 traffic quality 207
Smart Probe Data 288 Transmission Control Protocol (TCP) 262
smart vehicle management system (SVMS) 20 travel network 242
SNMP manager 317 TSPs (Telematics Service Providers) 2
social networking 253
source-to-destination path 207, 208 U
sparse RVC (SRVC) 231 Ubiquiti wireless cards 339
spatiotemporary coordination 220 Unwired network 341, 342, 345
SPER protocol 259 URVC (Ubiquitous Roadside-to-Vehicle Com-
SRVC system 231, 234, 236 munication) 235
start-of-frame (SOF) 4 Usage Data Records (UDR) 158
state-of-the-art 266 USB-to-serial link 345
STMP 271, 272, 274, 275, 280, 281, 283, User Agent Client (UAC) 162
284, 285, 286, 287, 289 User Agent Server (UAS) 162
Street RAndom Waypoint (STRAW) 365
sub-network 320, 321, 322, 323, 324, 325,
326, 328
410
Index
V WAVE management messages 90, 91, 97

WAVE multi-channel operations 90, 91
VADD protocol 211 WAVE network services 91, 94, 95
VANET 185, 190, 193, 196, 202, 204 WAVE resource management 91
VANET embedded system 291 WAVE short message protocol (WSMP)
VANET environment 364, 365, 366 92, 95, 96
VANET protocol design 355 WAVE short message (WSM) 91, 94
VANET simulations 355, 356, 357, 366 WAVE systems 185
VANETs technologies 183 WBEM-based Remote Management 310
Variable Message Sign (VMS) 54 WBEM-based Remote Vehicular Network
variable valve timing (VVT) 71 Management 311, 326, 327
Vehicle Ad Hoc Networks (VANETs) 229, 230 WBEM protocols 311
vehicle dispatch server (VDS) 23 web-based communication protocol 167
vehicle management program (VMP) 20 web-based enterprise management (WBEM)
vehicle navigation system 131, 132, 133, 311, 322
134, 135, 136, 137, 139, 141 Weigh-in-Motion (WIM) 34
Vehicle Navigation System (VNS) 131 wide area network (WAN) 319
vehicle to infrastructure (V2I) 91, 229, 230 WiMAX 144, 145, 147, 148, 149, 150,
vehicle to infrastructure (V2I) communication 153, 154, 155, 156, 157, 158, 159
229, 230 WiMAX core network
vehicle-to-roadside 90 144, 149, 150, 155, 158
vehicle-to-vehicle 90 Windows Management Instrumentation (WMI)
vehicle to vehicle (V2V) 91, 229, 230 316
vehicle to vehicle (V2V) communication Wired Equivalent Privacy (WEP) 377
229, 230, 253 wireless access in vehicular environment
vehicle tracking system 139 (WAVE) 7, 90, 91
Vehicular ad hoc network (VANET) 136, 183, wireless ad-hoc network 371
206, 252, 356, 370 Wireless channels 337
Vehicular Communication (VC) system 369 wireless communication 20
vehicular environment 160 wireless mesh 331, 332, 333, 335, 340, 345
vehicular network application 15 Wireless mesh networks (WMN) 331, 332, 340
vehicular network environment 311, 312 wireless personal area network (WPAN) 75
vehicular network middleware 310 wireless sensor network (WSN) 74
vehicular networks 348, 365
vehicular router 24 X
very-large-scale integration (VLSI) 67
VeSOMAC allocation 190 x-by-wire 5
video-broadcasting service 167 XML document 168
Video on demand (VOD) 158 XML format 169, 175, 176
Virtual TMC 288 Z
voice over IP (VoIP) 84
voronoi-based mobility model 365, 368 ZigBee 74, 75, 76, 77, 78, 79, 80, 81, 82,
83, 84, 85, 86, 87, 88, 89
W ZigBee topology 81
WAN (Wide Area Network) 10, 12 zone of relevance (ZOR) 219, 233
WAVE Basic Service Set (WBSS) 95, 96
411

Telematics Wireless Architecture and Application (2010) GOOD Lecture

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Telematics Wireless Architecture and Application (2010) GOOD Lecture

Uploaded by

Copyright:

Available Formats

Telematics Communication

InformatIon scIence reference

Published in the United States of America by

Library of Congress Cataloging-in-Publication Data

British Cataloguing in Publication Data

Foreword ............................................................................................................................................ xix

Preface ................................................................................................................................................ xxi

Acknowledgment ............................................................................................................................... xxv

Compilation of References ............................................................................................................... 380

About the Contributors .................................................................................................................... 398

Index ................................................................................................................................................... 405

Foreword ............................................................................................................................................ xix

Preface ................................................................................................................................................ xxi

Acknowledgment ............................................................................................................................... xxv

Compilation of References ............................................................................................................... 380

About the Contributors .................................................................................................................... 398

Index ................................................................................................................................................... 405

Zsehong Tsai, Professor

TelemaTics communicaTion Technologies and Vehicular

inTroducTion by the convention center even before stepping

Figure 1. In-vehicle and out-vehicle network architecture

Figure 2. Three nodes connected through a CAN

Figure 3.CAN message frame

Figure 4. LIN bus

Figure 5. FlexRay backbone

Figure 6. OSEK/VDX in-vehicle network management architecture

application messages. The network by combining different wireless networks. Cur-

Figure 7. Vehicle safety communication examples

Figure 8. DSRC spectrum band

Figure 9. Mobile WiMAX network architecture

neXT-generaTion Vehicular suitable platforms that would provide the desired

Figure 10. The WiMAX overview of the architecture

Figure 11. A block diagram of the next generation vehicle network

inTroducTion the wireless optimized TCP, the wireless optimized

Figure 1. Open telematics service

Wireless vehicle safety

Figure 2. The portal framework overview

First, the low-level context information such as Vehicular neTwork

Figure 3. Road-look-ahead service

Figure 4. SunLabs’ concept of car software architecture

Figure 5. Management of internal and external communications

Figure 6. Smart vehicle management system prototype design

Figure 7. Architecture of application manager

Figure 8. Architecture of the OSEK/VDX and OSGi-based embedded software platform

Figure 9. OSGi-based remote vehicular service platform

Figure 10. Application deployment in TOPAZ

Figure 11. Mobile telematics computing

Figure 12. Telematics architecture for mobile robot

Figure 13. Overview of middleboxes

Figure 14. The agent-based framework using JINI technologies

Figure 15. Vehicular security architecture

inTroducTion of advanced communication, sensing, and infor-

Infrastructure and (2) Intelligent Vehicles. In Advanced Traffic Management

Figure 1. Advanced traffic management systems (Source:ATMS)

Figure 3. Changeable message sign (Source: http://www.flickr.com/photos/moogal/442796974/)

Figure 4. Dynamic bus information systems

Electronic Payment Systems (EPS)

Electronic Payment Systems or Electronic Toll

Figure 7. Commercial vehicle operations

Figure 10. Automatic image detection system to identify incidents

Figure 12. Vulnerable individual protection systems

Information Management and agencies and facilitate decision making. Users

communications standards the status of signals operated from another