JOURNAL OF NETWORKS, VOL. 9, NO.

7, JULY 2014

1819

An Energy-efficient Opportunistic Multicast

Routing Protocol in Mobile Wireless Sensor
Networks
Shaojie Wen, Zhenyu Zhang*, Wenzhong Yang, and Siquan Hou
School of Information Science and Engineering, Xinjiang University, Urumqi 830046 China
*Corresponding author, Email: wsj_tian@163.com, zhangzhenyu@xju.edu.cn, ywz_xy@163.com,
xyz_hg@foxmail.com

AbstractThe
paper
presents
energy-efficiency
opportunistic multicast routing protocol (E-OMRP) for the
multicast energy consumption minimization problem in
mobile wireless sensor networks. The protocol divides the
network into grids, so each node determines their own
coordinates according to the grid. The nodes only need to
know the topology of their own grid, instead of the topology
of the entire network. In order to better represent the
impact of the movement on transmission, the nodes in the
same grid determine the priority in light of the transmission
delay factor and expected transmission cost. Then the relay
nodes use the opportunistic routing to send messages to the
next hop based on the order of priority. When the
destinations receive a message from relay nodes, they select
the optimal path in accordance with the transmission delay
and links cost, and transmission acknowledgement to source
along the optimal path selected. After the source has
received all the acknowledgements from different
destinations, it uses the method of path aggregation to
aggregate all optimal paths selected. If all nodes in the
current grid transmission messages failure, the neighbor
grids are in charge of retransmission. The results of
simulation show that, compared with traditional multicast
routing, E-OMRP is effective in reducing the consumption
of links cost, so that it improves the link reliability and
reduces the delay.
Index TermsMobile Wireless Sensor Network; Energy
Efficiency; Multicast; Opportunistic Routing; Path
Aggregation

I.

INTRODUCTION

Multicast has become an important means of

communication in the wireless sensor networks
(WSNs) ,in terms of task allocation and target query [1]
which has more advantages than unicast and broadcast.
Since the sensor nodes are deployed in remote, sparsely
populated areas, and energy constrained, so after the
nodes are deployed, recharging more difficult, so
improving the efficiency of each multicast transmission
link to minimize the energy consumption of nodes in the
network, which will not only extend the network work
lifetime, but at the same time reduces the costs spent on
the sensor hardware devices. So the research work on
how to increase the energy efficiency [2] of wireless
sensor networks is always a hot research area.
2014 ACADEMY PUBLISHER
doi:10.4304/jnw.9.7.1819-1827

In general, there are many ways to improve the energy

efficiency of nodes. One way is to decrease the total
communication traffic by using the data aggregation, data
compression, data prediction and duty cycling techniques
[3-5]. The second solution is to design effective network
protocols to increase the network lifetime. The first case
from the hardware point of view to reduce energy
consumption, while the second case from the viewpoint
of software to improve the energy efficiency. In order to
solve the reliable problem of wireless channel
transmission, MIT has proposed a new routing scheme in
recent years- opportunistic routing [6]. Traditional
wireless ad hoc network routing protocol use pre-defined
routing [7-8]: before the end to end data transmission, the
network establishes a transmission link between source
and destinations, and when each packet forwards, firstly
it will select one next hop node, and then forward packet.
When the links quality is poor, the frequent data
retransmission will consume a lot of bandwidth and
energy of the relay nodes, at the same time, reduces the
throughput of the network. The opportunistic routing can
solve the problems mentioned above, which take
advantage of the broadcast nature of wireless signals [911], through the competition of multiple potential relay
nodes to select the optimal next hop relay. Literature [12]
describes the different opportunistic routing, each method
designs the routing protocol with different network
metrics to ensure reliable transmission and reducing the
links cost. In order to better achieve the purpose of
energy efficiency, Changhun Bae et al. [13] [14] uses
mathematical methods between energy and delay, energy
and bandwidth obtain the optimal value, to find the
minimum energy value transmission required. B.
Manimozhi et al. [15] compares the different metrics in
WSNs. Dubois FH et al. [16] calculate the minimum links
cost based on opportunistic routing to achieve the
purpose of energy efficient.
In this paper, we present energy-efficiency
opportunistic multicast routing protocol (E-OMRP) to
minimize the transmission costs of each link. Firstly we
divide the network into grids [17], each node can
determine their coordinates based on the grid coordinate.
Nodes only need to know the current grid topology,
rather than the entire network topology. In order to make

1820

JOURNAL OF NETWORKS, VOL. 9, NO. 7, JULY 2014

the data transmission more reliable, we use opportunistic

routing to complete the transmission between links. The
protocol selects several next hop relay nodes, and each
relay node has an opportunity to transmit the messages to
next hop. When the wireless environment is unstable, it
can greatly improve the transmission reliable. Selecting
the appropriate set of relay nodes, not only reduce signal
interference between each other nodes, but also improve
the network throughput and reduce the links cost. In the
mobile wireless sensor networks, when all the relay nodes
in current grid transmission messages failure, the
neighbor grids are responsible for retransmission, and
combined with opportunistic routing and coordination
mechanisms [18-19], it can be good to solve the problems
of links interruption.
Major contributions of this paper as follow:
1 the nodes determine the priority in light of the
transmission delay factor and expected transmission cost,
where the nodes speed as a parameter to calculate
transmission delay factor, which can ensure that nodes in
a mobile situation, also can choose the optimal
forwarding set.
2 Combining opportunistic routing and collaboration,
the nodes in grid transmission the message to the next
hop using the opportunistic routing, when the messages
delivery failure, the current grid neighbors are
responsible for retransmission. It performs well in the
case of link state greatly change and guarantee the
reliability of messages transmission.
3 To obtain the purpose of improving the energy
efficiency, the paper adopts hops and links cost to
calculate path status, and selects optimal path based on
the value calculated. When the cost is the same, the
system will choose the path of least hops to be
responsible for subsequent transmission. When the
number of hops is same, the system selects the path
which the cost is least. The method not only ensures the
least cost, but also minimizes the number of hops.
4 The paper presents a method of path aggregation, to
reduce the number of links and minimize the number of
the working nodes in the network, which can improve the
average energy efficiency of nodes and prolong the life
time of network.
Section II gives some related work and problem
description. Section III defines a model for energy
consumption and describes the protocols under
consideration. Section IV presents results of the
simulation. Sections V discusses conclusions and future
directions.
II.

RELATED WORK AND PROBLEM DESCRIPTION

In WSNs and Mobile Ad hoc, multicast protocols

proposed typically use the tree structure as the network
topology [20-22]. For example, the paper [23] uses the
Takahashi-Matsuyama heuristic to establish steiner tree
for multicast protocol. In addition, the multicast
algorithms rely on the routing table to establish and
maintain multicast tree [24]. Tree-based routing protocols
can reduce the total number of hops and save bandwidth,
but when the network topology dynamically changes,

2014 ACADEMY PUBLISHER

protocols performance poorly in terms of throughput and

robustness. In order to overcome the weakness of tree
network, the literature [25-26] propose grid-based routing
protocol which establish multiple trees between source
and destinations, so the relay nodes can transmit
messages to each destinations along multiple paths. In the
period of messages transmission, if one of links arise
congestion or failure, the system can select other paths to
transmit the messages to destinations, which increases the
robustness and throughput of network. The literature [2728] uses the idea of grids for multicast, the nodes in
networks calculate the optimal path from source to
destinations based on certain routing metrics (for example:
bandwidth, throughput, jitter etc.). However, in treebased and grid protocols, the nodes need to create and
maintain routing information in the relay nodes, which
increase the links cost and network delay. In locationbased multicast [29-30], the messages transmission is
only based on the node location information. Each node
in the network only knows the multicast nodes location
and the local topology rather than the global topology,
which can reduce the number of beacons transmission in
network, at the same time, it also can reduce interference
between signals. Literature [31] has proposed a receiversbased stateless multicast method, where the nodes do not
maintain states information, only according to the
location information of destinations in the network for
transmitting
messages.
Literature
[32-33]
use
opportunistic routing for multicast, and combined with
network coding, effectively improve network throughput,
increasing the reliability of message transmission.
Different multicast protocols complete the data
transmission with different network topology and routing
algorithm. Although they have their own advantages and
disadvantages, in certain circumstances, can well satisfy
the need of messages transmission. Now we describe the
question: Suppose s represents source node, D
represents the set of destinations, d i is i th node of the
D , the path from s to d i expressed as P =(s, n1 , n2 ,
n3 nk , d i ), the total costs of path P represents C p . s
delivers message to d i along the path P, the delay time is
t , in the case of the reliability ascertain, obtaining a
balance between link costs and transmission delay. In this
paper we use the hop count to approximate the
transmission delay minimizing the cost meanwhile in
order to ensure that the hop count of path is least, and
vice versa.
III.

ENERGY EFFICIENCY OPPORTUNISTIC MULTICAST

ROUTING PROTOCOL

A. The Network Topology Description

We use digraph G ( E,V , C ( X , Y )) to simulate
network, in which E represents the links between the
grids, treating all the nodes in the single grid as a super
node V , C ( X , Y ) is costs that the messages transmitted
from the grid X to the next hop set Y . Assuming the grid
gi M , M represents the set of all the grids in network,

JOURNAL OF NETWORKS, VOL. 9, NO. 7, JULY 2014

1821

the set of all nodes within the grid is expressed as vgi , in

g i each node vngi Vgi . The edge between vngi within g i

and v p gi1 within gi 1 represents e(vngi , v pgi1 ) , and any
pair of nodes inside g i and gi 1 constitute the set of
edges
that
expressed
as
where
Egi , gi1 ,
e(vngi , v pgi1 ) Egi , gi1 . FLgi is the set of forwarding nodes

from the grid g i to its one hop neighbor grids. If the

current grid is g i , then the neighbor set in its one hop
transmission range represents N ( gi ) .

Figure 1. The selection of the optimal forwarding set

B. The Selection of Optimal Forwarding Set

In the routing protocol, the relay nodes set is very
important to the performance of the entire network. If the
selected relay set is optimal, which can reduce the
transmission delay, increase links reliable and improve
network throughput.
Based on the above several aspects, we use the concept
of social network subgroup [34] to select the optimal
relay nodes set. In opportunistic routing, the nodes in
forwarding set must locate in each others monitoring
range, this can ensure that each node can know whether
the messages successful transmission, based on the
feature, we use the subgroup cohesion degree [35]to
decide whether the set selected is the best.
We assume that there are two nodes ( X , Y ), on the
path from s to one of destinations, where X is upstream
node of Y . All nodes between X and Y , if they are
closer to Y than X be selected as candidate forwarding
set, and constitute a candidate forwarding set with X and
Y , also known as subgroup. Candidate forwarding set
can be expressed as S n , ni and n j are different nodes in

C. Optimal Route Selection

1) Energy Model
According to the relationship between communication
distance ( d 2 ) and distance threshold ( d 4 ), there are tow
channel models as free space and multi-path fading
respectively. If the communication distance is greater
than the threshold ( d d0 ), the system uses the multipath fading channel model, otherwise adopts the free
space model. Based on the nodes energy consumption
and the communication model, the initial energy of each
node is E n ( E n >0), the energy consumption of nodes to
send and receive k bit data are respectively expressed as
[36] [37]:

set, where ni , n j S n . If there is an edge between each

where Et and Er are the energy consumption of the

circuit to transmit and receive per bit respectively, when
the circuit is determined, Et is a constant and Eamp is the
power coefficient of the amplifier circuit, the energy
consumption of amplifier relate to communication
distance and bit error rate. d is the physical distance
between the sending node and the receiving node. n is
the channel loss index on the path, the value of which is
generally 2 or 4, in this paper we make n 2 , k is the
bit rate of the wireless signals. From the formula (1) we
can see, the energy consumption of transmission and
receive is proportional to data size, sending more data,
the more energy consumption. On the other hand, the
energy consumption of transmission also is proportional
to distance d , the farther the distance, the more energy
consumption. In this paper, we only consider the impact
of distance on the cost of nodes.
2) Path Metrics Calculation
In order to simplify the problem, we use the energy
consumption level of the nodes to approximate cost
metric. When a node receives a message from its
upstream node, the cost of nodes consume is equal to the
energy of nodes to receive messages consume, the cost of
transmitting nodes consumption is equal to the energy

pair of nodes, then it will have Sn ( Sn 1) 2 in

subgroup, the graph of subgroup structure called a
complete graph. xni , n j show whether there is a edge exist
between ni and n j , if so, xni , n j is1,otherwise is 0.

ni Sn

n j Sn

xni , n j is the actual number of edges in

subgroup. The subgroup cohesion degree can be

expressed as:

ni Sn

n j Sn

xni , n j

Sn ( Sn 1) 2

(1)

When 1 , the selected set is optimal. Sop as the

optimal set and is empty at the beginning time. We
choose the node with closer to center region, and add to
set. Then select other node in S n , and judge whether exit
a edge with every node in Sop . If the condition holds, so
the node is available and add to Sop , this process can not
end until S n is empty. Figure 1 shows, the black are the
nodes which meet the conditions.

2014 ACADEMY PUBLISHER

k * Et k * Eamp * d 2 , d d 0

ETx (k , d )
k * E k * E * d 4 , d d
t
amp
0

ERx (k ) k * Er

(2)

(3)

1822

JOURNAL OF NETWORKS, VOL. 9, NO. 7, JULY 2014

spent to send messages. As shown in Figure 2, assuming

that the central node X in the grid g i transmits
messages to the node Y in the grid gi 1 , where the
packet loss rate is eX and the transmission cost is wX ,Y .
The k th node of forwarding set receive messages from
the upstream node X , where the packet loss rate is pk
and the transmission cost is wX , k , and transmission
messages to the downstream nodes from k node the
packet loss rate is ek and the cost is wk ,Y . The signal

We use the speed factor and expected transmission

count [9], to calculate the priority of nodes in forwarding
set, the expected cost is:
ECk

Ck
(1 pk ) (1 ek )

(8)

ECk is the priority parameter of forwarding nodes and

the smaller the value, the higher the priority of nodes.

received power is proportional to distance d [38],

therefore it can know that the links reliability is relation
to distance, so we use distance parameter to represent the
link volatility. The total cost of forwarding nodes in one
transmission is:
Ck wX , k wk ,Y wsensor

(4)
Figure 3. a express the direction vector from n to Y and b is the
velocity vector of n, is the angle between the two vectors.

where wsensor is the consumption cost of node to sense.

Then we compute the expected transmission costs

ETCgi that node X transmit messages to Y [39]. First
calculate the cost of consumption of each node in
forwarding set, the cost of consumption receive messages
from node X and send it to Y .
FL

ETCgi (1 eX )CX ( j 1 e j ).(1 ek )Ck (9)

k 1

k 2

The probability g i transmit message to the next hop

gi 1 is expressed as:

Egi , gi1 1 j 1 e j
FL

(10)

Figure 2. Different nodes transmission cost in same forwarding set

To simplify operation so gi j 1 e j , supposing the

FL

We assume that ngi is a candidate forwarding node in

g i and node Y in gi 1 , all nodes in the network at the

same speed v and the movement direction of nodes are
uncertainty. As shown in Figure 3, a express the direction
vector from ngi to Y and b express the velocity vector of
ngi , is the angle between the two vectors. We can

neighbors set of g i is N ( gi ) , the neighbors of which

receive messages correctly is denoted as S N ( gi ) ,
through collaboration between the grids to transmit
messages from g i to gi 1 , the probability and expected
number of transmissions is, respectively:
Et ( gi , gi 1 ) 1 (1 Egn , N ( gn ) )

calculate cost as

(11)

gn S

cos

ab
ab

(5)
ETX ( gi , gi 1 )

The distance from ngi to Y is d, after a unit of time,

the distance is:

dt d - v cos

(6)

In order to calculate the nodes transmission cost more

accurately, a speed factor is considered:

d v cos dt

d
d

2014 ACADEMY PUBLISHER

(7)

1
1 (1 gn )

(12)

g n S

According to equation (4) and (5) obtained the

transmission cost ( CTC ) of g i in collaboration with
neighbors as follows:
CTCgi Egi , gi1 * ETCgi

(1 E

S N ( gi ) g k S

g k 1 , gi 1

) *( Egk , gi1 ) * ETCgk

(13)

JOURNAL OF NETWORKS, VOL. 9, NO. 7, JULY 2014

1823

When k 1 , (1 Egk 1 , gi1 ) indicates the probability of

g i transmit messages failure. When g i transmit message

to gi 1 , in the case of regional collaboration, by the (7)
and (8) can calculate the total expected transmission cost
[40] expressed as:
ETCt ( gi , gi 1 ) ETX ( gi , gi 1 )* CTCgi

(14)

Definition 1: Let P=(s, g1 , g 2 , g 3 g i , d k ) is a path

that one of the multicast routing set , g i is ith path on
the path, the costs of routing path P expressed as C ( P ) ,
where the sum of link cost and the expect link costs of the
path P as follow:
C(P )

gi P

gi , N ( g i )

cg1 , N ( gi ) cg2 , N ( g2 ) ... cgk , N ( gk ) (15)

many factors, the volatility of the network link is

relatively large, the volatility will decrease the
transmission reliability of the hop-to-hop. Unreliable
links will lead to more nodes or more areas are
responsible for the task to deliver the message, when the
number of nodes and areas reach a certain value, the total
transmission cost will be greater than the sum of the
expected transmission costs. In the normal case, it is more
suitable for realistic scenario. The second condition for an
ideal state, in this environment the links is stable, the
reliability of message transmission is high, but this ideal
state is incompatible with the actual situation.
In order to find the optimal transmission path, this
paper proposes a link metric functions, which include two
parameters one of the parameters , the other parameters
. F is expressed as follows:

F f ( , )
ETC ( P )

ETC ( g , g

gi P

i 1

(16)

Assuming there is a path P between the source and

destination, when the data transmission rate is determined,
the message transmission delay between g i and gi 1 can
be known. There are k bit data are delivered from g i to
gi 1 , the transmission delay can be expressed as Di . On
the basis of formulation (11), we can calculate the
expected transmission delay and the total path delay:
Dexp Di ETX ( gi , gi 1 )
Dtpd

D ETX ( g , g

gi p

i 1

(17)
)

(18)

In the actual transmission, the delay of path P is DT ,

is delay factor which use the expected transmission
delay and actual transmission delay to calculate, with the
following formula to express:

DT Dtpd
Dtpd

(19)

When the source has multicast messages to send for

the first time, the path of transmission multicast messages
maybe not optimal for each destination. In order to
minimize the costs of subsequent message transmission,
we use the actual cost and expected transmission costs of
the links to calculate the optimal path.
According to (14) and (15) to calculate , because the
network is affected by many factors, the current status of
links depends on the conditions of their environment. To
Ht
simplify the formula, let
H best

C ( P ) ETC ( P ) *
ETC ( P )

(20)

The first case indicates that the network link status is

poor than the ideal state, in a real environment due to
2014 ACADEMY PUBLISHER

(21)

When the costs are same, the network may choose the
path with fewer delay time as the subsequent messages
transmission path.
D. Path Aggregation
Because the density of nodes deployed in sensor
networks is relatively large, each node in the network has
many neighbors. In the initial stage, the source need
choose an optimal path for each destination according to
the different routing metrics. The paths selected using this
method have redundancy greatly, the energy efficiency of
the entire network without considering yet. As shown in
Figure 4, before source C sends data to destinations A and
B, C need to select two paths for destinations, we assume
that the network selects optimal path in according to the
minimum delay, the optimal paths from C to A and B are
C-1-3-5-A, C-2-4-6-B respectively. So the data will walk
along the definitive path. It can be calculated easily, there
are nearly eighty percent of the network nodes in the
working state. From the energy efficiency point view, it is
not desirable. Now we consider this problem from a
different angle, assuming that A and B in the transmission
of node 7, if along the C-1-3-7- A and B or C-2-4-7- A
and B, then the working nodes in the network will reduce
about half, at the same time, it saves the energy of nodes
to a certain extent.
The analysis above is the idea and advantages of path
aggregation [11]. Next we will introduce the specific
methods of data aggregation. We extract the source and
destinations from Figure 4, so they can form an other
figure, as Figure 5, which only includes A, B and C. The
C are connected with A and B respectively, so the two
lines form a angle, expressed as . The transmission
distance of nodes as the radius to draw a circle, In twodimensional space, moving the circle, until it tangent to
two lines. There is a node n at the center of circle, which
is true node or a virtual, the path from source C to n can
be as optimal fusion path. From the traditional point of
view, the throughput of single path is low, and also prone
to cause the problem of single node failure. We do not
have to worry about these problems, because the
opportunistic routing used in the paper, which can solve

1824

JOURNAL OF NETWORKS, VOL. 9, NO. 7, JULY 2014

them perfectly. Making the circle tangent to two lines, in

order to guarantee that when the network environment is
not stable, making the data transmission quickly and
efficiently recover.

Figure 4. In the process of data transmission, if the protocol does not

use the data aggregation, there will be two paths in the network,
expressed as solid lines, after considering the path aggregation, only
single path exit in the network, as dotted lines represent

Figure 5. represents the intersection angle between the edges CA

and CB, node n is the center of circle, which can be true node or virtual

Now we briefly introduce the concept of degree of

aggregation, short for DA. Assuming that the hops from C
to A and B are h1 and h2 respectively, and to n is h3 , so
DA can be calculated as follow.
DA

h3
Max{h1 , h2 }

(22)

Through the calculation we can know, when > 900 ,

DA=0, 900 , DA>0. So the range of DA is 0 to 1.
E. Protocol Description
Assuming all grid coordinates information have been
saved in the node's cache. In the network initialization
phase, all nodes in the grids sense their current
coordinates and according to the center position of the
grids to determine whether it is in the forwarding set. If
the node is in the forwarding area, and it calculates the
expected transmission costs and broadcast to its
neighbors according to the center coordinates of the
upstream and downstream transmission with the packet
contains (Node ID, the grid ID, expected transmission
cost value). Forwarding node are not in the area only sent
the packets contains (Node ID, where the grid ID) to the
neighbors. When the nodes receive data packets from
different grids, directly discard. By now, the nodes in the
forward set have a list that contain the identifier of all the
nodes in grids and it would remove a node not in
forwarding set, afterwards they determine their own
priority based on the expected cost item. The nodes
which have the highest priority send query messages to
the neighbor grids, and the information of the neighbor
grids are stored in the cache of each forwarding node,
which include expected transmission costs from its
2014 ACADEMY PUBLISHER

neighbor grids to current grid, such network initialization

is complete. Before the source sends data to the
destinations, it firstly obtains the optimal paths after
aggregating according to the coordination of the
destinations, and then starts the process of the data
transmission. Because the paper regards the grid as the
network topology, so realizes the path aggregation with
the grid as a unit yet.
In order to meet parameter metrics, we will modify
several items in packet header. As the topology treated as
the basic unit in grids in the paper, the data packet is
transmitted between grids, so the entries ForwardNum
in the packet header is not suitable for the proposed
protocol requirements, and replaced as GridNum . In
addition, the corresponding ForwardList entry will be
replaced by GridList . In the data frame format, add
expected transmission costs and the actual transmission
costs of message transmission, and extracted as ETCost
and RTCost . After destination nodes have received the
data packets from a plurality of neighbors, extract
ETCost and RTCost using equation (18) to calculate .
Then based on the coordinates of source and destinations,
it would calculate expected transmission hops, using the
message header TTL entries according to the formula (19)
to calculate . Finally, using equation (20) to calculate
F , and select the path with the smallest path as the
optimal path, transmission the acknowledgement to the
source node along the reverse of the optimal path.
IV.

EXPERIMENTAL RESULTS AND ANALYSIS

We simulate E-OMRP on the C++ platform, in

networks a lot of nodes are randomly deployed in the area
of 1000m*1000m, the transmission rate of nodes are the
same as 1Mb/s, the packet size is 512kb, the wireless
signal transmission range is 147m. In order to illustrate
the performance of the E-OMRP protocol, the protocol
compared with MAODV [41], the average energy
consumption of each hops in the network and the average
transmission node number per hops with the different
number of nodes and different speeds. We assume that
the external interference effect on the link transmission
reliability is small, the energy consumption of nodes
monitoring is not included in the cost.

Figure 6. The average energy consumption of nodes in case of the

different number of nodes

As shown from figure 6, E-OMRP protocol scales well

with respect to the energy efficiency. When the number is

JOURNAL OF NETWORKS, VOL. 9, NO. 7, JULY 2014

small, the available relay nodes are less, the nodes which
are responsible for transmission are not theoretically
optimal, the energy consumption of two methods are
more than the number of nodes is large. As the number of
nodes increase, the cooperation chance among neighbors
can be fully taken advantage of leading to the delivery
ratio gain compared to the MAODV protocols. When the
number of nodes changes, average energy consumption
of the nodes in MAODV is more than the method we
proposed. Because the method proposed in the paper uses
the opportunistic routing and cooperation mechanism.
When a node or a grid transmission failure, which need
not to retransmission from the source node, which can
ensure that the network choose the optimal link to
transmission messages, rather than the links with low
transmission rate. When the nodes in the network are
relatively small, there are less paths can be aggregated, in
this case, the paths aggregation could not work well.
While the number of nodes is larger, the paths aggregated
are more. Figure 7 shows that the energy consumption of
the nodes with different speeds, the movement of nodes
may lead to links failure. At the beginning, the delivery
ratios slightly decrease as the velocity increases, It is
because nodes are more likely to outrun communication
ranges. The links in MAODV are more sensitive to the
moving speeds of nodes, the average energy consumption
of nodes increase with the speeds. When the speed
increases, E-OMRP performs better than MAODV.
Because the speed as a parameter to calculate the priority
of nodes, which can ensure the reliability of the
transmission.

1825

Figure 9. The number of nodes per hops transmission need with

different speeds

Figure 8 and 9 consider both the number and moving

speeds of nodes, we compare E-OMRP and MAODV in
terms of the average number of nodes per hop
transmission. We can see from tow figures, when the
nodes is less in the network, the two methods could not
successfully transmission the message to the next hop
with high probability, one side there are few relay nodes
to select, on the other hand, in the transmission process,
there are not suitable next hop to transmission messages.
In this paper, we use transmission delay and cost as the
metrics to select the optimal path, the average number of
nodes per hop transmission can reflect the quality and
reliability of links, and use path aggregation to reduce the
number of links and the interference among nodes. As
can be seen from the figures, when the number of nodes
is less, E-OMRP uses opportunistic routing and select the
optimal node set to transmit messages, which is better
than MAODV in terms of transmission reliability without
the use of opportunistic routing. When the MAODV
protocols slightly perform poorly due to breakage of links
resulting from the node mobility, the breakages may
make nodes that are far away from the source difficult to
receive the data packets. Our method performs well,
because when the links dynamic are large, single grid
transmission success rate is low, the cooperation
mechanism will makes single hop transmission becomes
more reliable.
V.

Figure 7. The average energy consumption of nodes with different

speeds

Figure 8. The number of nodes per hops transmission need in case of

the different number of nodes

2014 ACADEMY PUBLISHER

CONCLUSIONS

In this paper, we have proposed E-OMRP for the

energy efficiency to illustrate the energy efficiency
problem, and use the opportunistic routing and
cooperation as well as path aggregation to transmit the
messages from source to the destinations. Relay nodes set
use the opportunistic routing to transmit the messages,
which can ensure the reliability of transmission and
improve the throughput of network. When the nodes
quickly move, the links is unstable, so the current grid
needs the help of their neighbors to delivering the
message to the next hop. Adopting transmission delay
and links cost as metrics to select optimal path, which
may ensure that when the delay time is certain, the
network will select optimal path with minimum cost.
Although opportunistic routing is good enough to solve
the link instability problems, however, to make it give
full play to its strengths in multicast, which needs for

1826

JOURNAL OF NETWORKS, VOL. 9, NO. 7, JULY 2014

further research. The method in this paper only considers

network layer, but data link layer and physical layer are
taken less consideration, and it can run well in network
layer, this does not represent that it can meet the
requirements of the other layers, and the research of
cross-layer protocol will be the focus of our future work.
In the future work, we plan to design more efficient
forwarding set selection algorithm and add the reliability
to our E-OMRP protocol instead of using simple
acknowledge along the optimal path selected.
ACKNOWLEDGMENT
The work is supported by National Natural Science
Foundation of China (Project No. 61262089 and
No.61262087). The authors wish to thank many referees
for their suggestions on improving this paper.
REFERENCES
[1] I. F. Akyildiz, W. Su, Y. Sankarasubramaniam, and E.
Cayirci Wireless sensor networks: a survey, Computer
Networks, vol. 38(4), Mar. 2002, pp.393-422.
[2] A. Giuseppe, M. Conti, M. Di Francesco, and P. Andrea,
Energy conservation in wireless sensor networks: A
survey, Ad Hoc Networks, vol. 7(3), May. 2009, pp.537568.
[3] W. Ye, J. Heidemann, and Estrin D, Medium Access
Control With Coordinated Adaptive Sleeping for Wireless
Sensor Networks, IEEE/ACM Transaction, vol. 12(3), Jun.
2004, pp.493-506.
[4] J. J. Xiao, A. Ribeiro, and Z. Q. Luo, Distributed
compression-estimation using wireless sensor networks,
IEEE Signal Processing Magazine, vol. 23(4), July. 2006,
pp. 27-41.
[5] M. C. Vuran, I. F. Akyildiz, Spatial correlation-based
collaborative medium access control in wireless sensor
networks, IEEE/ACM Transactions on Networking, vol.
14(2), Apr. 2006, pp.316329.
[6] B. Sanjit, R. Morris, Opportunistic routing in multi-hop
wireless networks, ACM SIGCOMM Computer
Communication Review, vol. 34(1), 2004, pp.6974.
[7] C. Perkins, E. Royer, Ad-Hoc on-demand distance vector
routing, Proc. of the IEEE WMCSA99, In Mobile
Computing Systems and Applications, IEEE Computer
Society Press, Feb. 1999, pp.90100.
[8] C. Intanagonwiwat, R. Govindan, and D. Estrin, Directed
diffusion: a scalable and robust communication paradigm
for sensor networks, Proc. of the ACM Mobi-Com 2000,
6th annual international conference on Mobile computing
and networking, ACM PRESS, Aug. 2000, pp.5667.
[9] S. Biswas, R. Morris, ExOR: opportunistic routing in
multi-hop wireless networks, Proc. of the ACM
SIGCOMM 2005, Applications, technologies, architectures,
and protocols for computer communications, ACM Press,
Oct. 2005, pp.133143.
[10] G. Jean-Marie, Z. Ruifeng, and K. Jaffr`es-Runser, G.
Claire, Energy, latency and capacity trade-offs in wireless
multi-hop networks, pro. 2010 IEEE 21st international
symposium, personal and mobile radio communications,
IEEE PRESS, Sep. 2010, pp. 2757-2762.
[11] Z. Zhong, J. Wang, and S. Nelakuditi, Opportunistic anypath forwarding in multi-hop wireless mesh networks,
Doctoral dissertation, University of South Carolina, 2006.

2014 ACADEMY PUBLISHER

[12] T. Ke, Z. Baoxian, M. Jian, and Y. Zheng, The

opportunistic routing in wireless multi hop networks,
Journal of software, vol. 21(10), 2010, pp. 2542-2553.
[13] B. Changhun, and W. E. Stark, End-to-end energy
bandwidth tradeoff in multihop wireless networks, IEEE
Transactions on Information Theory, Vol. 55(9), Sep. 2009.
[14] B. Changhun, and W. E. Stark, Minimum energy-per-bit
multi-hop wireless networks, Pro. Forty-Sixth Annual
Allerton, Communication, Control, and Computing, IEEE
PRESS, Sep. 2008, pp. 23-26.
[15] B. Manimozhi, and B. Santhi. Comparison of different
performance measures of routing protocols in WSN,
International Journal of Engineering and Technology
(IJET), Vol. 5(1) Feb-Mar 2013, pp.208-214.
[16] H.F. Dubois, M. Grossglauser, and M. Vetterli, Least-cost
opportunistic routing, Pro. 2007 Allerton, Communication,
Control and Computing, vol. 11, 2007, pp. 13-39.
[17] K. Chakrabarty, S. Iyengar, H. Qi, and E. Cho, Grid
coverage for surveillance and target location in distributed
sensor networks, IEEE Transactions On Computers, vol.
51(12), Dec. 2002, pp. 1448-1453.
[18] M. Kurth, A. Zubow, and R. Jens-Peter, Cooperative
opportunistic routing using transmit diversity in wireless
mesh networks, Pro. IEEE INFOCOM 2008, 27th
conference on Computer Communications, IEEE PRESS,
April 2008, pp. 1310-1318.
[19] H. Haifeng, and Y. Zhen. Collaborative opportunistic
routing in wireless sensor networks, Journal Of
Communications, vol. 30(8), Aug. 2009, pp. 116-126.
[20] L. SungJu, S. William, and G. Mario. On-Demand
multicast routing protocol in multi-hop wireless mobile
networks, Mobile Networks Applications, vol. 7(6), Dec.
2002, pp. 441-453.
[21] J. Garcia-Luna-Aceves and E. Madruga, A multicast
routing protocol for ad-hoc networks, Proc. IEEE
INFOCOM, Eighteenth Annual Joint Conference of the
IEEE Computer and Communications Societies
(INROCOM 99), IEEE PRESS, vol. 2, Mar. 1999, pp. 784792.
[22] R. Vaishampayan and J. Garcia-Luna-Aceves, Efficient
and robust multicast routing in mobile ad hoc networks,
Proc. IEEE Intl Conf, Mobile Ad-Hoc and Sensor Systems,
IEEE PRESS, Oct. 2004, pp. 304-313.
[23] O. Akihito, I. Takeshi, and M. Akira, BAM branch
aggregation multicast for wireless sensor networks, Proc.
IEEE International Conference. Mobile Ad hoc and Sensor
Systems, IEEE Press Nov. 2005, pp.10-363.
[24] C.-C. Chiang, M. Gerla, and L. Zhang, Shared tree
wireless network multicast, Proc. Sixth Intl Conf,
Computer Communications and Networks, Sept. 1997, pp.
28-33.
[25] B. Jeff, C. Tracy, T. Vishy, Mesh-based geo-cast routing
protocols in an ad hoc network, Pro. IPDPS01, the 15th
International Parallel & Distributed Proceeding
Symposium, Apr. 2001, pp.184.
[26] C. Tracy, and Y. Liub, An adaptive mesh-based protocol
for geo-cast routing, Journal of Parallel and Distributed
Computing, vol. 63(2), Feb. 2003, pp. 196213.
[27] H. Dhillon, and H. Q Ngo, CQMP: a mesh-based
multicast routing protocol with consolidated query
packets, Pro. Wireless Communications and Networking
Conference, IEEE 2005, vol. 4, Mar. 2005, pp.2168-2174.
[28] Z. Guokai, W. Chen, and L. Xiao, Grid multicast: an
Energy-efficient multicast algorithm for wireless sensor
networks, Proc. Fourth International Conference on
Networked Sensing Systems (INSS07), IEEE Press, Jun.
2007, pp. 267-274.

JOURNAL OF NETWORKS, VOL. 9, NO. 7, JULY 2014

[29] A. Sanchez, and M. Ruiz, Bandwidth-Efficient

geographic multicast routing protocol for wireless sensor
networks, IEEE Sensors Journal, vol. 7(5), May. 2007, pp.
627-636.
[30] W. Shi-bo, and K. Selcuk, GMP: Distributed geographic
multicast routing in wireless sensor networks, Proc. the
26th IEEE International Conference on Distributed
Computing Systems, IEEE Press, 2006, pp. 1-9.
[31] C. Hsiang, Z. Yuqun, and D. Ilker, Stateless multicast
protocol for ad hoc networks, IEEE Transactions On
Mobile Computing, vol. 11(2), Feb. 2012, pp. 240-253.
[32] Y. Wenzhong , H. Chuanhe, and Z. Zhenyu, An Efficient
Opportunistic Multicast Routing for Mobile Ad Hoc
Networks, Proc. 2nd International Conference on Future
Computer and Communication, IEEE Press, May. 2010,
pp. 316-321.
[33] K. Abdallah, I. M. Khalil, and W. Jie, Distributed network
coding-based opportunistic routing for multicast, Pro.
MobiHoc '12 Proceedings of the thirteenth ACM
international symposium on Mobile Ad Hoc Networking
and Computing, 2012, pp. 115-124.
[34] Bock, Husain, An adaptation of holzingers bcoefficients
for the analysis of sociometric data, Sociometry, May.
1950, pp.146-153.
[35] Alba, A graph-theoretic definition of a socio-metric
clique, Journal of Mathematical Sociology, 1973, pp:
113-126.
[36] L. Marie Feeney, An energy consumption model for
performance analysis of routing protocols for mobile ad
hoc networks, Mobile Networks and Applications, vol.
6(3), Jun. 2001, pp. 239249.
[37] W. Alice, B. Wendi, and S. Amit, Energy-scalable
protocols for battery-operated micro-Sensor networks,
Journal of VLSI Signal Processing systems for signal,
image and video technology, vol. 29(3), Nov. 2001, pp.
223237.
[38] H. You, L. Zhiyun, A geographically opportunistic
routing protocol used in mobile wireless sensor networks,
2012 9th IEE International Conference on Networking,
Sensing and Control (ICNSC), 2012, pp.216-221.
[39] H. DuboisFerriere, M. Grossglauser, and M. Vetterli,
Valuable detours: Least-cost any-path routing,
IEEE/ACM Transaction on Networking, vol. 19 (2), April.
2011, pp. 333-346.
[40] M. Xufei, T. Shaojie, X. Xiaohua, and L. Xiangyang,
Energy-Efficient opportunistic routing in wireless sensor
networks, IEEE Transactions On Parallel and Distributed
Systems, vol. 22(11), Nov. 2011, pp. 1934-1942.
[41] E.M. Royer and C.E. Perkins, Multicast operation of the
ad-hoc on-demand distance vector routing protocol, Proc.
of the 5th annual ACM/IEEE international conference on

2014 ACADEMY PUBLISHER

1827

Mobile computing and networking. ACM, 1999, pp. 207218.

Shaojie Wen was born in Henan, China.

He is now a graduate student of School
of Information Science and Engineering,
Xinjiang University Urumqi China, and
will finish school in July, 2014. His
interests include opportunistic network
and wireless sensor network multicast.

Zhenyu Zhang was born in Datong,

China. He received the Masters degree
in Electronic Science and Technology
University. He is now associate
professor and assistant Dean of School
of Information Science and Engineering,
Xinjiang University of Urumqi, China.
His interests include opportunistic
network, p2p network and wireless
sensor network.
Prof. Zhang is a member of the CCF
Wenzhong Yang was born in Urumqi,
China. He received the Ph.D degree in
information security from the Wuhan
University of Hubei province in 2011.
He is now associate professor and
assistant Director of School of
Information Science and Engineering,
Xinjiang University of Urumqi, China.
His interests include p2p network and
wireless sensor network.
Siquan Hou was born in Henan, China.
He is now a graduate student of School
of Information Science and Engineering,
Xinjiang University Urumqi China, and
will finish school in July, 2014. His
interests include opportunistic network
and social network.