You are on page 1of 4

2008 International Symposium on Computer Science and Computational Technology

A MAC Protocol Based on S-MAC for Power Asymmetric WSN Network

Yun Li, Hao Jia, Zhan-jun Liu, Zhi Ren, Guo-jun Li


Research Laboratory of Mobile Communications, Chongqing University of Posts and Telecommunications, Chongqing 400065, China e-mail: tthh1130@sina.com
AbstractThe communications are interrupted by the hidden terminals in the traditional MAC protocols when the nodes power is asymmetric in WSN networks. On the basis of SMAC the improved protocol is proposed. Node judges whether the power asymmetric nodes are occurred in the vicinity of it according to the link state with neighborhood nodes. Then neighborhood nodes broadcast Bandwidth Reservation (BRES) message intelligently after overhearing CTS. After receiving BRES the high power nodes will inhibit their own transmissions for the duration. So the improved protocol avoids the interruption of hidden terminals, reduces the collisions, decreases the delay and increases throughput. The simulation proves the performance of improved protocol is better than S-MAC in power asymmetric WSN networks. Keywords-WSN, MAC, broadcasting BRES message, hidden terminals

I.

INTRODUCTION

In recent years, wireless sensor network technology becomes an emerging field in wide range of applications such as environment monitoring, battlefield surveillance, monitor of human physiological data, and etc [1]. The low cost sensors with limited power sources can be deployed in differing places for differing usages, and they are able to sense temperature, fluid levels, humidity and other attributes. It is usually not cost effective to recharge the batteries of the sensors deployed on distinct environments; sensors are designed to carry limited irreplaceable power sources. Therefore, sensor network lifetime will rely on the corresponding batteries of sensor nodes. Collisions avoiding is a basic mission of MAC protocol, it decides the node in the network when, how visiting medium and send out data. A good MAC protocol can avoid collisions effectively; reduce retransmission of data and save energy. Currently the MAC protocols of WSN are almost based on TDMA [2-4] and based on competition [5-6]. In the TDMA-based protocols, SMACS [2] sends the invitation of connection periodically using controlling channel, consults with the adjacent nodes of the data channel connecting slot and frequency. Nodes sleep in the slot of disconnection. According to the energy level PACT [3] chooses some special node to transmit data packet, the other nodes is closed to save energy. TRMA [4] exchanges nodes message and schedule with neighborhood nodes, uses distributed algorithm to determine the node to send data in the coming slot, the other nodes maintain sleep. S-MAC [5] is a singlechannel MAC protocol based on competition. Nodes
978-0-7695-3498-5/08 $25.00 2008 IEEE DOI 10.1109/ISCSCT.2008.263 299

overhear the channel periodically and compete to send or receive data packet, in the absence of competition nodes sleep. In the basis of S-MAC T-MAC [6] change the duty cycle adaptively, in accordance with the network traffic nodes change the time of competition, educe the idle listening greatly. These traditional protocols that are typically designed with an implicit assumption that links are bi-directional. But in real WSN networks the energy of each node only relates to operation measures and communications measures. Because of different workload, each node has different energy. In such a heterogeneous network, different nodes are likely to have different power capabilities and thus, are likely to transmit with different power levels. This, in turn, leads to possible link asymmetry wherein the transmission of a high power node is received (or is sensed) by a lower power node whereas the high power node cannot sense the transmissions by the low power node. Thus, the low power nodes are hidden from high power nodes. This increases the number of collisions that are experienced by low power communications.
Range of 3

Range of 2

3
Range of 0 RTS CTS

CTS

Range of 1

Range of 4

Figure 1. Hidden terminal problem due to link asymmetry

This effect is depicted in Figure 1. The RTS/CTS exchange between the two low power nodes 0 and 1 is not overheard by node 3 since node 3 is not within the sensing or interference range of these communicating nodes. Thus, during the data exchange between nodes 0 and 1, node 3 could possibly begin its own transmission, for example node 3 sends data to node 4, thereby causing a collision at node 1. From the above we can find that the traditional MAC protocols can not avoid collisions effectively in power asymmetric WSN network, so designing a MAC protocol to adapt the power asymmetric WSN network is necessary.

This paper presents an improved protocol on the basis of SMAC. The improved protocol can reduce collision effectively in power asymmetric WSN networks. II. IMPROVEMENT OF S-MAC PROTOCOL S-MAC protocol adopts the RTS/CTS mechanism to address the hidden terminal problem. There is a duration field in each transmitted packet that indicates how long the remaining transmission will be. So if a node receives a packet destined to another node, it knows how long it has to keep silent. The node records this value in a variable called the network allocation vector (NAV) and sets a timer for it. Every time when the NAV timer fires, the node decrements the NAV value until it reaches zero. When a node has data to send, it first looks at the NAV. If its value is not zero, the node determines that the medium is busy. However, as shown in Figure 1, the above-mentioned asymmetrical power, only the RTS/CTS mechanism can not eliminate the possibility of collisions completely in power asymmetric WSN network. Improved protocol increases broadcasting a new control message on the basis of the RTS/CTS, named BRES control message Bandwidth Reservation. This control message is derived from the CTS message, which contains duration field. The BRES message format is similar to the RTS message format except that that the frame control field has an additional attributes: the originator address field which contains the MAC address of the node that initiated the communication (RTS sender). These fields are used for the detection of duplicate BRES messages that may be received. The improved protocol adds an attribute B in CTS message. When node overhears CTS, it examines whether B in CTS message is 1. If yes, node broadcasts BRES message; if not, node only throw the CTS and set its NAV. The modified CTS and BRES packet formats are shown in Figure 2.

Figure 3. Flow chart of awareness for the link asymmetric nodes

Figure 2. Message structure

In the improved protocol we require that each node maintain link-state information. We can know which nodes are symmetric or asymmetric from the link-state information. Towards this, each node broadcasts a hello message periodically, which contains ID of sender. When node receives hello message, it replies chello message, which contains ID of sender and ID of destination. Node preserves the senders ID of chello message in a list, named symmetric table. And each node preserves the senders ID of hello message in a list, named asymmetric table. When the hello message period is finished, node compares the symmetric table and the asymmetric table. if two tables are different, the link asymmetric nodes are occurred in the vicinity of it. Node adds correlative cue message in CTS message. The flow chart is shown in Figure 3.

With improved protocol, when two nodes exchange RTS/CTS, in the scope of broadcasting RTS/CTS the neighbors overhear CTS message, according to B in CTS message node determines whether it needs to broadcast BRES control message. If B in CTS message equals 1, node broadcasts BRES message; if B in CTS message equals 0, node just set NAV as S-MAC. Nodes broadcast BRES message after waiting a random number (between 0 and 6) of short interframe space (sifs) units. This is because each neighborhood node hears the CTS at the same time, neighborhood nodes determine to send BRES messages at the same time, and there will be collisions because of BRES message. Waiting for a random number (between 0 and 6) of short interframe space (sifs) units before broadcasting the BRES message is to minimize collisions due to multiple simultaneous BRES broadcasts from the neighbors that hear the same CTS message. Via broadcasting BRES control message the link asymmetric nodes which are occurred in the vicinity of node know about the ongoing RTS/CTS exchange between the sender and the receiver so that they would inhibit their own transmissions for the duration. The goal is to inform high power nodes that can potentially initiate transmissions while the low power communication is in progress about the communication, so that they defer their transmissions until after the completion of the low power data transfer. On the opposite direction when nodes receive CTS message, nodes need to send data wait for an extended interframe space (eifs) unit for preventing low power nodes from sending data before the broadcast of BRES. The complete RTS/CTS-based data transfer sequence with our scheme is depicted in Figure 4.

300

SIFS

6* SIFS

node 0 and node 4 is cbr0, the data flow between node 3 and node 4 is cbr1. The parameters of cbr0 are the same with cbr1, which include data packet size, transmission rate and transmission interval. They start to transmit data at 10s. We choose the Simulation environment as shown in Table 1. In all our figures, the parameter along the abscissa indicates the interval of transmission.
TABLE I. SIMULATION ENVIRONMENT 1 Mbps 40% 1000 bytes 0.5 Mbps 600 s 1 ms 50 ms 5 ms

EIFS

Bandwidth Duty Cycle Packet Size Rate Simulation Time Slot EIFS SIFS

Figure 4. The modified reservation scheme

The broadcast of BRES control message is depicted in Figure 5. The RTS/CTS exchange between nodes 0 and 1 is not overheard by high power node 3. Node 2 overhears the CTS, examines the B in CTS message equals 1, so node 2 broadcasts BRES message. High power node 3 receives BRES message, notices the transmission of node 1, then sets NAV to inhibit its own transmissions to node 4 for the duration specified, so avoiding collision at node 1.

Figure 5. Propagating BRES messages.

The RTS/CTS exchange between the two low power nodes 0 and 1 is not overheard by node 3 since node 3 is not within the sensing or interference range of these communicating nodes, during the data exchange between nodes 0 and 1, node 3 could possibly begin its own transmission, thereby causing a collision at node 1 with SMAC protocol. Collisions at low power node 1 instigate rediscovery of routes, TCP retransmissions and timeouts and can result in the wastage of wireless capacity and poor utilization efficiency. With improved protocol, the high power node 3 is able to sense the RTS/CTS exchange between the two low power nodes 0 and 1; it would inhibit its own transmissions for the duration, so that the improved protocol reduces the collisions at node 1. We are able to reduce the number of collisions as compared to S-MAC protocol in the Figure 6.

III.

SIMULATION

We study the performance of the proposed techniques using the event driven packet network simulator, ns2, specifically the version ns2.31 [7]. The simulated network is deployed in a square region. The total number of nodes in the network is fixed at 5. We chose nodes that can transmit with either of two transmission power levels, 7.575*10-3 W or 3.03*10-2 W for simulating a heterogeneous WSN network, two of the chosen nodes were high power nodes and the other three were low power nodes. The power of nodes 0, 1 and 2 is 7.575*10-3 W; the power of nodes 3 and 4 is 3.03*10-2 W. When data packets are generated at nodes 0 and 3, they are destined for the only sink node 4. The nodes topology of our simulation is shown in Figure 5: node 4 is the sink node; node 0 and node 3 are the source nodes. Node 1 is the neighbor node of node 0; Node 0 and node 2 are the neighbor nodes of node 1; Node 1, node 3 and node 4 are the neighbor nodes of node 2; Node 2 and node 4 are the neighbor nodes of node 3; Node 2 and node 3 are the neighbor nodes of node 4. The transmission of node 3 is received (or is sensed) by node 1 whereas node 3 cannot sense the transmissions by node 1.The data flow between

Figure 6. Collisions at low power node 1

In Figure 7, we show the improvements seen by the nodes in terms of throughput with the use of the improved protocol compared to S-MAC. The throughput of the low power node 1 is improved. The reduction in number of retransmissions by low power nodes reduces the overall contention for wireless medium and as mentioned earlier, the

301

collisions are alleviated. We observe an overall improvement in network throughput as compared to S-MAC.

decreased. At lower interval, the collisions increase and the improved protocol tends to provide benefits. If we further increase the interval, the possibility of collisions reduces and therefore the benefits seen due to our schemes decrease. IV. CONCLUSIONS In this paper we consider WSN networks that consist of nodes that have different transmit power capabilities. Real networks are likely to consist of such nodes. S-MAC protocol performs poorly in such networks. The primary reason for this is the fact that high power nodes could potentially be beyond the range of low power communicating nodes and might therefore initiate transmissions so that collide with low power packets. We propose effective ways to reduce the collision. These techniques are based on using an intelligent broadcast scheme to propagate the BRES message which is derived from CTS. We show that using our scheme we can improve the Data Success Rate of low power nodes and alleviate the Link Failures between Low Power nodes as compared to SMAC protocol. We conclude that our MAC layer framework offers a simple and effective protocol for medium access control in power heterogeneous WSN networks. ACKNOWLEDGMENT This paper is supported by the National Science Foundation of China (60702055), Program for New Century Excellent Talents in University (NCET), the Science and Technology Research Project of Chongqing Municipal Education Commission of China (KJ070521). REFERENCES
[1] AKYILDIZ 1F, WEILIAN SU, SANKARASUBRAMNIAM Y. A survey on sensor networks[J]. Communications Magazine, IEEE, 2002, 40(8):102-114. SOHRABI K, GAO J, AILAWADHI V. Protocols for SelfOrganization of a Wireless Sensor Network[J]. IEEE Personal Communications, 2000, 7(5):16-27. PEI G, CHIEN C. Low power TDMA in Large Wireless Sensor Networks[J]. Military Communications Conference (MILCOM,01). Communications for Network-Centric Operations: Creating the message Force. 1EEE, 2001, 1(4):28-31. RAJENDRAN VOBRACZKA KGARCIA-LUNA-ACEVES JJ Energy-Efficient Collision-Free Medium Access Control for Wireless Sensor Networks[A] The First ACM Conference on Embedded Networked Sensor Systems [C] Los AngelesCA USANovember 2003. Wei Ye, John Heidemann, Deborah Estrin. An energy-efficient MAC protocol for wireless sensor networks[C]. IEEE Proc Infocom, 2002. 1567-1576. DAM TV, LANGENDOEN K. An Adaptive Energy-Efficient MAC Protocol for Wireless Sensor Networks[A]. The First ACM Conference on Embedded Networked Sensor Systems [C]. Los Angeles. CA. USA, November 2003. NS2: Network Simulator. http://www.isi.edu/nsnam/n.

Figure 7. Throughput of low power node 1 and sink node

From Figure 8, we observe a significant improvement in terms of delay between low power node 0 and sink node 4. The RTS/CTS exchange between the two low power nodes 0 and 1 is not overheard by node 3 since node 3 is not within the sensing or interference range of these communicating nodes, during the data exchange between nodes 0 and 1, node 3 could possibly begin its own transmission, thereby causing a collision at node 1 with S-MAC protocol. Collisions instigate data retransmissions. With improved protocol, the high power node 3 is able to sense the RTS/CTS exchange between the two low power nodes 0 and 1 and would inhibit its own transmissions for the duration, so that improved protocol reduces the collisions at low power node 1. The reduction in number of retransmissions by low power nodes reduces delay compared to S-MAC.

[2]

[3]

[4]

[5]

[6] Figure 8. End to end delay

Notice from Figures 7 and 8 that at high interval, the performance improvements are not significant. This is because, with the larger interval the number of transmission is reduced, this implies that possibility of collisions is

[7]

302

You might also like