Content uploaded by Phouthone Vongpasith
Author content
All content in this area was uploaded by Phouthone Vongpasith on Sep 03, 2024
Content may be subject to copyright.
Phouthone Vongpasith et al., International Journal of Advance Research, Ideas and Innovations in Technology.
© 2017, www.IJARIIT.com All Rights Reserved Page | 1175
ISSN: 2454-132X
Impact factor: 4.295
(Volume3, Issue2)
Available online at www.ijariit.com
Movement Direction Algorithm for Geographic
Routing Protocol in Vehicular Networks
Vongpasith Phouthone *
College of Computer Science and Electronic Engineering
Hunan University, Changsha, 410082, China
lb2011019@hnu.edu.cn
Zhu Xiao
College of Computer Science and Electronic Engineering
Hunan University, Changsha, 410082, China
zhxiao@hnu.edu.cn
Dong Wang
College of Computer Science and Electronic Engineering
Hunan University, Changsha, China
wangd@hnu.edu.cn
Vincent Havyarimana
Department of Applied Sciences, Ecole Narmale Superieure,
6983, Bujumbura, Burundi
havincent14@hnu.edu.cn
Abstract—Vehicular ad hoc networks (VANETs) are special type of the wireless network in which communication
through other intermediate vehicles on the road. Due to high mobility of vehicles, the design of an efficient routing
protocol is one of key issues, which has been considered a challenging problem to deal with such dynamic network in
VANETs. The greedy-perimeter stateless routing (GPSR) routing protocol uses the simple greedy forwarding based only
on the position information which may fail to find a neighbor closer to the destination than itself. In this paper, we
propose a movement direction algorithm under GPSR. It comprehensively takes into account the velocity vector
information and underlying link expiration time to recover a local maximum. The simulation outcomes in varying
scenarios show that the proposed algorithm enhances the packet delivery and reduces the end-to-end delay, compared
to the existing geographic routing algorithms.
Keywords— GPSR, Local maximum, Movement direction, Next-hop, VANETs
I. INTRODUCTION
Vehicular ad hoc networks (VANETs) are special type of the wireless network in which communication take place
through wireless links mounted on each vehicle [1]. Each vehicle within VANETs acts as both, the participant and router
of the network. The vehicles communicate through other intermediate vehicle on the road that lies within their own
transmission range [2]. The movement direction of vehicles is along roadways, and their mobility is restrained by traffic
policies, such as traffic light signals, speed constraints, and road/traffic conditions [3]. The basic target of VANETs is to
increase safety of road users and comfort of passengers. VANETs has the unique and critical characteristics, such as: the
vehicles may join or leave within one another transmission ranges abruptly or gradually [4], the network topology may
change rapidly and unpredictably, the established wireless links between the vehicles may break. As a result of this
mobility, the challenges of VANETs will become more serious and the performance of routing protocols could be greatly
affected. Due to high mobility and uneven distribution of vehicles, the design of an efficient routing protocol is one of key
issues, which has been considered a challenging problem to deal with such dynamic network scenarios [5]. In VANETs,
the routing strategy is whether we select an appropriate or undesirable next-hop vehicle from the neighbouring set. The
performance still not perform well, if the routing cannot find an optimal next-hop vehicle, as which is essential entity for
delivering of the packets from the source to the destination. Therefore, the performance of routing relies on the most
suitable next-hop selection mechanism for data delivery among vehicles.
To address these specific requirements of VANETs, greedy perimeter stateless routing (GPSR) [6] is one of the
geographic routing protocols that can use the local topology information to find correct new routes quickly and it has been
actually designed for dynamic network scenarios [2]. It requires neither regular exchange of the routing information nor
broadcast flooding to route requests [3]. It does not need to store routing information; thus it is very effective as a dynamic
Phouthone Vongpasith et al., International Journal of Advance Research, Ideas and Innovations in Technology.
© 2017, www.IJARIIT.com All Rights Reserved Page | 1176
network. To select a next-hop neighbor, GPSR uses the simple greedy forwarding based only on the position information.
It may choose the next-hop that is the closest to the destination but moving in the opposite direction of the destination
vehicle. Therefore, it misses out on some suitable candidates to forward a packet; Moreover, [7] the greedy forwarding
may encounter the local maximum or local optimum problem, where the current forwarder is closer to the destination than
all its neighbors and the destination still not reachable by one hop communication [8]. For example a vehicle in front may
not have any neighbor nearer to it than the final destination vehicle [9]. In such a case, GPSR uses perimeter phase to
overcome the situation when the greedy phase meets a local maximum situation [1]. It uses the right hand rule to search
routes at the boundary in a direction to route the packet along the perimeter of the local maximum region in a clockwise
direction [6]. In VANETs, with the right-hand rule, there may be large number of vehicles on the left side of the road, and
the packet may be seen to go further and further away from the target, until the life cycle of the last vehicle is reduced to
the end, and then the packet is dropped [1]. The performance of data delivery is degraded, as well as the increase in network
delay.
To explore routing features of VANETs, in this paper we propose a movement direction algorithm under the GPSR
routing by considering the velocity vector of vehicles. First, instead of perimeter routing to recover the local maximum,
we select the optimal next-hop vehicle by considering the number of vehicles that move in the direction of the destination
vehicle. Then, we consider the vector projection of each candidate next-hop into consideration to select the optimal next-
hop node. Otherwise, we take into account the link-reliability between each neighbor and the destination into consideration
based on the link expiration time (LET) information. In our work, we have performed extensive simulation of VANETs
based on urban scenario traces generated from VanetMobiSim [10] as the input to the proposed model based on NS-2
network simulation [11]. We have studied the impact of important factors such as number of vehicles and maximum speed
of vehicles for comparing its performance with those of the existing geographic routing algorithms. Simulation results
clearly show that the proposed is viable and can significantly enhance the packet delivery ratio, reduce end-to-end delay,
compared to others.
The remainder of this paper is organized as follows. Related work is reviewed in Section 2. Section 3 introduces the
movement direction model. Then, the simulation setup and results are shown in Section 4. Finally, we conclude the paper
in Section 5. II. RELATED WORK
In VANETs, many algorithms under the geographic routing strategies have been designed to forward the packet from
the source to the destination vehicles [12], [13], [14], [15], [16]. GPSR is one of the geographic routing that uses only the
neighbor’s position information to forward the data packets [6]. In greedy forwarding, upon receiving a data packet with
the destination’s position information, the source selects a neighbor that is closest to the destination and forward the data
packets to that neighbor [17]. During the process of the greedy forwarding, a local maximum occurs, the vehicle would
switch to perimeter routing that attempts to route the packet along the perimeter of the local maximum region in a clockwise
direction [4]. If during perimeter routing, the packet reaches a vehicle that is closer to the destination than the vehicle at
which the routing entered into perimeter mode, the vehicle would resume the greedy forwarding of the received a packet
[18]. However, planarization of the graph is very difficult because of the mobility of the VANETs. The recovery strategy
of GPSR is inefficient and time consuming especially given the highly dynamic nature of VANETs [9]. The routing with
perimeter forwarding in GPSR may lead to wrong directions, and there are too many hops for the packet to be transmitted
to the destination which can lead to the packet loss and delay. To improve the GPSR routing protocol, the simplification
of perimeter forwarding which is based on GPSR but takes the position information, the speeds and directions of movement
of the neighboring vehicles into accounts, to predict the future positions before to forward a packet, and to ensure the
reliability of the routing [12]. By considering stochastic characteristics of the distance and speed of vehicles for computing
the link-reliability that link with reliability factor greater than a given threshold alone is selected as a next-hop neighbor,
when constructing a route from the source to the destination [14]. The performance improvement over the conventional
GPSR protocol in terms of the packet delivery ratio and link failure rate is significantly reduced; however the delay slightly
increases as compared to the conventional GPSR. Due to the potentially large number of next-hop neighbors, a next-hop
selection scheme in [19] uses the optimal stopping theory to choose a suitable next-hop neighbor, while in [20] uses the
future position of each neighbor and then selects neighbor vehicle nearest to the next intersection based on predictive
location. The predictive location estimates the future location of a vehicle based on its history location and velocity [4],
[21]. To predict the probability of the link-reliability between two vehicles, dynamic properties are used in the movement
states of vehicles. In the stable state, the mean velocity of vehicles is stable, otherwise, in the unstable stable, the velocity
of vehicles is unstable with the acceleration or deceleration [22]. In movement prediction based routing (MOPR) [23],
before selecting of a next-hop vehicle each vehicle estimates the link-reliability in its transmission range based on the
movement information such as velocity and direction. Then, MOPR will select the next-hop vehicle with the highest the
link-reliability to forward a data packet. However, as the vehicle does not know the real location of the target vehicle, it is
impossible for it to evaluate the prediction error [4].
Knowledge of the link-reliability is essential for the design of the geographic routing protocols. Recently, there have
been certain attempts to analyse the link-reliability in VANETs [22], [24], [25]. In link state aware geographic routing
protocol, a routing metric called expected one-transmission advance (EOA) is contrived to improve the greedy forwarding
algorithm by diminishing transmission failures [24]. The EOA and link-reliability in [25] are measured using the enhanced
the expected transmission count (ETX) metric which is obtained with the assistance of the information (e.g., position and
velocity) in the hello massage. The calculation of ETX is modified to adapt to the high mobility of the network vehicles.
Since the ETX metric depends highly on the value of the hello interval and window size [24].
Phouthone Vongpasith et al., International Journal of Advance Research, Ideas and Innovations in Technology.
© 2017, www.IJARIIT.com All Rights Reserved Page | 1177
Researchers have proposed next-hop selection strategies [19], [26], [27]. The selection of next-hop vehicle algorithms
resulting in with weak forwarding links. These weak forwarding links may degrade the performance of routing. For stable
and reliable routing, the selection of next-hop vehicle is very crucial task. Link and node based metrics have been identified
and used in the next-hop vehicle selection algorithm [27]. The node based metrics are localized metrics related to choose
a next-hop vehicle whereas link based metrics are depicting the quality of the link between two vehicles. To improve the
GPSR routing, a next-hop selection mechanism based on a weighted function which consists of the link-reliability between
the source and neighbour vehicles, distance between neighbour and the destination, movement direction angle of vehicles
is studied [26]. However, the performance of the proposed protocol is better in some situations. The simplest next-hop
selection strategy is to greedily select a neighbour vehicle with the highest geographical progress toward the destination
vehicle as a next-hop [22]. A next-hop selection strategy based on the length of the buffer takes not only the distance
between next-hop and the destination vehicle [15] but also the available length of the next-hop vehicle buffer into
consideration. Thereby reducing the time delay as well as the packet loss which caused by bigger waiting time than the
retransmission delay.
All of these routing protocols have proposed various algorithms to improve the performance of the geographic routing
in VANETs. A few of them have made an improvement to recover from the local optimum situation. In this paper, we will
propose a new recovery method for the greedy forwarding to improve the performance of the GPSR routing protocol in
VANETs. III. THE PROPOSED ALGORITHM
In Fig. 1 the source S initializes the data communication by sending the data packet to the destination vehicle D, the
traditional greedy forwarding of the GPSR protocol defines vehicle N as a next-hop, since vehicle N is clearly the closest
neighbor to the destination vehicle D. Here, the vehicle N is closer to the destination vehicle D than its neighbours. In this
case, the GPSR protocol declares N as a local maximum to D. The perimeter forwarding mode, where the right-handle rule
is used to find a perimeter around the line ND, and the packets are then forwarded from N along the path of the arrows.
This forwarding may lead to wrong directions, and there are too many hops for the packet to be transmitted to the destination
which can lead to the packet loss and delay.
This section shows the development of movement direction model in next-hop selection mechanism. We analyze the
velocity vectors of candidate and the destination nodes. According to the next-hop selection design principles, it requires
to comprehensively consider various performances in next-hop selection decision-making, so as to achieve performance
balance of routing protocols in aspects of the packet delivery ratio and end-to-end delay. We present a movement direction
model. It improves and optimizes the traditional GPSR protocol and mainly takes into account benefits of velocity vector
and LET on next-hop selection. During the process of the greedy forwarding, a local maximum occurs, as shown in Fig. 1,
the vehicle would not switch to perimeter routing, but firstly select a group of neighbor list based on the movement direction
information. It means the node whose movement trend is most close to the destination is selected as a candidate node, and
then we utilize the vector projection of each candidate next-hop into consideration to select the optimal next-hop node.
Otherwise, we take into account the link reliability between each neighbor and the destination into consideration based on
the LET information.
Fig. 1 The local maximum of GPSR
A. Assumption
To begin, the following is assumed to be reasonable. Each vehicle in the network can obtain the information of its own
and that of neighbor. The vehicles are equipped with global positioning system (GPS) devices that can provide information
about vehicle speed, direction, and position, are shown in Table 1. The destination’s information is added in the data packet
header in order to be available at the source and neighboring vehicles as shown in Table 2.
TABLE 1
INFORMATION IN THE HELLO PACKET
ID
Position
Velocity vector
Speed
Time
TABLE 2
INFORMATION IN THE DATA PACKET HEADER
Phouthone Vongpasith et al., International Journal of Advance Research, Ideas and Innovations in Technology.
© 2017, www.IJARIIT.com All Rights Reserved Page | 1178
ID
Mode
Source’s
information
Destination’s
information
B. Movement Direction Model
Throughout this paper, we will denoteas a neighbor of the source node S for the next-hop forwarder,
whereasi=1,2,...n is the number of total neighboring nodes. D denotes the destination node.
Vx,VyDenotes the
velocity vector of node, similarly,
Vx,Vydenotes the velocity vector of the destination node D. denotes the angle
between the movement direction ofand D. and a scalar product of two vectors
and
can be calculated by Equations
1 and 2, respectively.
cos
.
(1)
.
=VxVx+VyVy (2)
The vector projection of a vector onto a non zero vector
is a vector parallel to
defined as OH
, WhereOH
is a
scalar, called the scalar projection of
onto
when <π
, that is defined as Equation 3.
OH
.
(3)
According to Equations 1 and 3, when the neighboring node is moving in the direction of the destinationcosØ ,
Otherwise, the vehicle is moving in the opposite direction of the destination, cosØ . Obviously, when the value
ofcosØis larger, the scalar projectionOH
is larger, as shown in Fig. 2.
Fig. 2 Movement direction model
C. Link Reliability Model
LET is a function of time, it considers the position, velocity vector and speed for estimating the lifetime of
communication between two nodes. As shown in Fig. 1, if we consider two nodes and D with a transmission range of R,
the predicted LET is obtained as Equation 4 [28].
LET= +cadbcab+cd
+c (4)
a=scosθcos (5)
b=x (6)
c=ssinθsinθ (7)
d=y (8)
Where, respectively,y,, anddenotes the position, speed, and movement direction angle with respect to x-
axis of neighboring node . Similarly, ,y,, anddenotes the position, speed, and movement direction angle with
respect to x-axis of the destination node D.
D. Next-hop Selection Algorithm
In this subsection, the proposed next-hop selection algorithm based on the above movement direction and link-reliability
model is described. The algorithm is designed by considering the following two cases of recovery models.
In case 1, when neighboring nodes are moving in the direction of the destination, according to movement direction angle
of each arriving neighboring node and the destination node ranges in [0,
[, (cosØ , according to Equation 1). The
neighboring node which has the highest scalar projection and movestoward to the destination, is selected as an optimal
next-hop to forward the data packet from the source node to the destination, as shown in Fig. 2.
In case 2, when the movement direction angle between the neighboring nodes and destination node ranges in [
, [,
(cosØ ). In this case, we take the link-reliability of each neighbor and the destination node into consideration. The
Phouthone Vongpasith et al., International Journal of Advance Research, Ideas and Innovations in Technology.
© 2017, www.IJARIIT.com All Rights Reserved Page | 1179
selection of the best link reliability is made based on the longest lifetime of communication between two nods. Another
key concept is, when a local maximum occurs, all of the neighbors are moving in the opposite or orthogonal direction of
the destination, the neighboring node which has the highest value of LET takes the highest priority to be an optimal next-
hop node.
Now, let us turn our attention back to the local maximum of GPSR. In Fig. 1, for each neighboring node we let the
scalar projection of
onto a nonzero vector
beOH
, if we choose as the next-hop forwarder, and the estimation values
ofOH
will be obtained. Let node has the highest value ofOH
among the (the total number of neighboring nodes)
values. Then, nodewill be selected to be the next-hop forwarder and the packets will be forwarded to it. The packet
relaying will be repeated hop by hop according to the above process until the data packets will reach the destination. The
detailed next-hop selection algorithm of the proposed is given in Algorithm 1. This technique allows an decrease of
delivered delay to the destination. Because the packets doesn’t attempts to route along the perimeter of the local maximum
region that may lead to wrong directions.
Algorithm 1 Next-hop selection algorithm
OH
the initial value of scalar projection
the initial value of LET
next-hop -1
while do
/*the candidate node i
/* whereasi=1,2,...n is the number of total neighboring nodes
OH
the scalar projection of
onto
LET between the candidate node and destination D
if
then
/*the candidate node moves in the direction of the destination D
if OH
<OH
then
OH
OH
/* loop to find the candidate node with the highest scalar projection
next-hop
end if
else if <then
next-hop
end if
end while
return next-hop IV. SIMULATION SET UP AND RESULTS
A. Simulation Set up
To determine the performance of the proposed algorithm, we carried out simulations in NS-2 of version 2.35 [11]. We
simulated the protocols in a 1500 m x 1500 m rectangle scenario that is generated by the vehicular mobility model generator
VanetMobiSim [10], an open source program which can generate more realistic vehicular mobility for NS-2. Vehicles
move according to the intelligent driver model with lane changing model support smart intersection management: they
slow down and stop at intersections, or act according to traffic lights, if present. Also, vehicles are able to change lane and
perform over takings in presence of multi-lane roads. The propagation model used in the simulation is the Two-way Ground
model and transmission range of each vehicle is set to 250 m. There are 75 to 135 vehicles randomly distributed initially
on the roads. Once the simulation begins, each vehicle moves at a speed ranging from 5 to 10, 15, 20, and 25 m/s. The
simulation parameters are summarized in Table 3. TABLE 3
SIMULATION PARAMETERS
Parameters
Specifications
Simulation area
1500 m x 1500 m
Transmission range
250 m
Number of nodes
75, 95, 115, 135
Transport protocol
Transmission Control Protocol
(TCP)
Simulation time
250 s
Maximum Speed of
nodes
10, 15, 20 and 25 m/s
B. Simulation Results
Phouthone Vongpasith et al., International Journal of Advance Research, Ideas and Innovations in Technology.
© 2017, www.IJARIIT.com All Rights Reserved Page | 1180
This section presents simulation results and describes our observations. We compared the performance of the proposed
algorithm to MOPR [23], EOA [24], and GPSR [6]. We conducted extensive simulations based on impacts of vehicular
traces with the following performance metrics [3]:
the packet delivery ratio represents the ratio of the packets delivered to the destinations to those generated by the
sources.
the average end-to-end delay is defined as the average amount of time spent by the transmission of a packet that is
successfully delivered from the source to the destination.
1) Impact of Number of Nodes:
Fig. 3 shows the packet delivery ratio for varying the number of vehicles with maximum speed 15 m/s. An increase in
the number of vehicles slightly decreases the packet delivery ratio. The decrease comes from the fact that the routing
topology becomes more dense and unstable when network density increases which makes the network connectivity
unstable. Since, EOA metric incorporates the geographic distance and link-reliability and MOPR determines the most
stable link-reliability in terms of communication lifetime by selecting the most stable optimal next-hop, which are more
successful to deliver the data packets to the destination than the traditional greedy forwarding. Thus, the packet delivery
ratio achieved by EOA and MOPR is higher than GPSR. We note that the packet delivery ratio of the proposed approach
outperforms MOPR, EOA, and GPSR, which successfully deliver approximately 81.38% of the data packets, while MOPR,
EOA and GPSR deliver approximately 72.86%, 72.24% and 69.22%. This is because, the optimal next-hop selection in the
proposed approach takes into account the movement direction model to avoid the forwarding of a packet to a wrong next-
hop vehicle, which could result in the loss of the data packets. Otherwise, it recovers the local maximum situation with
higher LET, compared with that based on the link-reliability and perimeter recovery mode in the EOA, MOPR, and GPSR.
Fig. 3 Packet delivery ratio with different number of vehicles.
Fig. 4 shows the average end-to-end delay for varying the number of vehicles with maximum speed 15 m/s. An
increase in network density increases the chances to meet an appropriate next-hop and decreases the distance of vehicles
in which the packet delay should be reduced from the source to destination vehicles. Thus, the average end-to-end delay
of GPSR and proposed decreases with an increase in the numbers of vehicles. On the contrary, when the density of
vehicles is sparse, the connectivity of the network topology affects the end-to-end delay. Thus, if the number of vehicles
increases from 75 to 95, the average end-to-end delay quickly increases for EOA and MOPR. In the proposed scheme,
the selection process takes less time to find the optimal next-hop, by considering the vehicle which moves toward to the
destination, thus, the packet takes less time to reach the destination vehicle, it shows lower average end-to-end delay
values than others. Compared with the EOA, MOPR, and GPSR, the proposed decreases the average end-to-end delay by
48.14%,51.13%, and 54.60% on average, respectively.
Phouthone Vongpasith et al., International Journal of Advance Research, Ideas and Innovations in Technology.
© 2017, www.IJARIIT.com All Rights Reserved Page | 1181
Fig. 4 Average end-to-end delay with different number of vehicles.
2) Impact of Maximum Speed of Vehicles:
In Figs. 5 and 6 we study the impact of maximum speed of vehicles. Fig. 5 shows the performance of the packet delivery
ratio for varying maximum speed of vehicles. The packet delivery ratio increases for all the three routing protocols: EOA,
MOPR, and GPSR because an increase in the maximum speed of vehicles resulting in an increase in the opportunities for
the data packets to find out the optimal next-hop and improves the connectivity of network, which reduces the packet loss.
GPSR selects the next-hop only by a simple greedy forwarding technique with the geographic distance based on the position
information; thus, a data packet may enter a local maximum and recover through a link with poor quality, resulting in low
packet delivery ratio. Compared with the EOA, MOPR and GPSR, the proposed considers the link-reliability based on the
velocity vector and LET information to select the optimal next-hop when a local maximum occurs, which outperforms the
other routing protocols. On average, the proposed increases the packet delivery ratio by 14.66%, 14.99%, and 15.96%,
respectively.
Fig. 5 Packet delivery ratio with different maximum speed of vehicles.
As can be seen from the results shown in Fig. 6, the average end-to-end delay performance of EOA, MOPR, and GPSR
dramatically increases, with respect to maximum speed of vehicles increases to 25 m/s. This is due to the highly dynamic
network topology and frequent changes cause of a high packet delay and disconnection issues. It is observed that when the
maximum speed is 25 m/s, for EOA, MOPR, and GPSR, the end-to-end delay reaches over 175 ms, 172 ms, and 200 ms;
in contrast, for the proposed, the delay nearly stays to 70 ms. The proposed shows less end-to-end delay by forwarding the
data packet through the optimal next-hop selection based on movement direction model which has high link-reliability.
Compared with EOA, MOPR, and GPSR, the proposed algorithm decreases the average end-to-end delay by 43.02%,
48.96%, and 59.42% on average, respectively.
Phouthone Vongpasith et al., International Journal of Advance Research, Ideas and Innovations in Technology.
© 2017, www.IJARIIT.com All Rights Reserved Page | 1182
Fig. 6 Average end-to-end delay with different maximum speed of vehicles.
V. CONCLUSIONS
In this paper, We incorporated the movement direction model and link reliability based on velocity vectors and speeds
of vehicles. To recover the local maximum and enhance the performance of the GPSR routing protocol, we take the benefit
of a scalar product and scalar projection, which considers the candidate node that moves toward to the destination.
Otherwise, we use the link-reliability, which considers the candidate node that has the highest value of LET. Finally, we
investigated a comprehensive set of effecting factors, such as number and maximum speed of nodes to compare the
proposed approach and existing algorithms. The simulation results reveal that the proposed approach can achieve a better
performance in terms of the packet delivery ratio, with an increase of 11.59%, 12.06%, and 14.59%, compared to EOA,
MOPR, and GPSR. In the case of average end-to-end delay, the proposed approach performed best and is, 45.85%, 50.05%,
and 57.01% lower than EOA, MOPR, and GPSR. Although the proposed algorithm performs better than the other existing
algorithms under many configurations, the proposed model is useful only for best effort services in sparse scenarios. In the
future work, we will aim to consider the high density network topology to further improve the packet delivery.
ACKNOWLEDGMENT
This work was partly supported by the National Natural Science Foundations of China (No. 61272061 and No.
61301148), Hunan Provincial Natural Science Foundation of China (No. 2016JJ3041), and Open Foundation of State Key
Lab of Integrated Services Networks of Xidian University (ISN17-14).
REFERENCES
[1] H. Huang and S. Zhang, “A routing algorithm based on dynamic forecast of vehicle speed and position in VANET,”
International Journal of Distributed Sensor Networks, vol. 9, pp. 1-9, 2013
[2] H. Ghafoor, I. Koo, and N. Gohar, “Neighboring and connectivity-aware routing in VANETs, The Scientific World
Journal, vol. 2014, pp. 1-10, 2014.
[3] I. Abbasi, M. Bilal, and S. A. Madani, “A traffic flow-oriented routing protocol for VANETs,” EURASIP Journal on
Wireless Communications and Networking, vol. 2014, pp. 1-14, 2014.
[4] Y. Zhu, Z. Li, and M. Li, “Geographic routing based on predictive locations in vehicular ad hoc networks,” EURASIP
Journal on Wireless Communications and Networking, vol. 2014, pp. 1-9, 2014.
[5] X. M. Zhang, K. H. Chen, and Ohters, “A street-centric routing protocol based on microtopology in vehicular ad hoc
networks,” IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, vol. 65, pp. 5680 - 5693, 2016.
[6] B. Karp and H. T. Kung, “GPSR: Greedy perimeter stateless routing for wireless networks,” in Proc. The Sixth Annual
ACM/IEEE International Conference on Mobile Computing and Networking (MobiCom 2000), Boston, MA, 2000,
pp. 243-254.
[7] A. Fonseca and T. Vazao, “Applicability of position-based routing for VANET in highways and urban environment,”
Journal of Network and Computer Applications, vol. 36, pp. 961-973, 2013.
[8] T. Darwish and K. A. Bakar, “Traffic aware routing in vehicular ad hoc networks: characteristics and challenges,”
Telecommun Syst, vol. 61, pp. 489 -513, 2016.
[9] J. Liu, J. Wan, and Others, “A survey on position-based routing for vehicular ad hoc networks,” Telecommun Syst,
vol. 62, pp. 15 -30, 2015.
[10] J. Hrri, M. Fiore, and Others, “VanetMobiSim: generating realistic mobility patterns for VANETs,” in Proc. The 3rd
ACM International Workshop on Vehicular Ad Hoc Networks (VANET’06), Los Angeles, USA, 2006.
[11] Network Simulator. [Online]. Available: http://www.isi.edu/nsnam/ns
[12] H. Tu, L. Peng, and Others, “GSPR-MV: a routing protocol based on motion vector for VANET,” in: IEEE
International Conference on Signal Processing (ICSP), HangZhou, China, 2014, pp. 2354-2359.
Phouthone Vongpasith et al., International Journal of Advance Research, Ideas and Innovations in Technology.
© 2017, www.IJARIIT.com All Rights Reserved Page | 1183
[13] S. B. Lahlah, F. Semchedine, and L. B. Medjkoune, “A position-based routing protocol for vehicular ad hoc networks
in a city environment,” in Proc. The International Conference on Advanced Wireless, Information, and
Communication Technologies (AWICT 2015), Tunisia, 2015, pp. 102-108.
[14] S. Shelly and A. V. Babu, “Link reliability based greedy perimeter stateless routing for vehicular ad hoc networks,”
Hindawi Publishing Corporation International Journal of Vehicular Technology, vol. 2015, pp. 1-16, 2015.
[15] T. Hu, M. Liwang and Others, “An enhanced GPSR routing protocol based on buffer length of nodes for the
congestion problem in VANETs,” in: Proc. The 10th International Conference on Computer Science & Education
(ICCSE 2015), Fitzwilliam College, Cambridge University, UK, 2015, pp. 416-419.
[16] K. Monika and A. Sharma, “A Deterministic GPSR Routing Protocol for VANET,” International Journal of Advance
Research , Ideas and Innovations in Technology, vol. 2, pp. 1-5, 2016.
[17] Balasubramani, L. Karthikeyan, and V. Deepalakshmi, “Comparison study on non-delay tolerant routing protocols in
vehicular networks,” in Proc. 2nd International Symposium on Big Data and Cloud Computing (ISBCC15), VIT
University, Chennai, 2015, pp. 252 – 257.
[18] L. Hu, Z. Ding, and H. Shi, “An improved GPSR routing strategy in VANET,” in Proc. The International Conference
on Wireless Communications, Networking and Mobile Computing (WiCOM), Shanghai, China, 2012, pp. 1-4.
[19] J. Qian, T. Jing, and Others, “A next-hop selection scheme providing long path lifetime in VANETs,” in Proc. IEEE
26th International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC): Mobile and
Wireless Networks, Hong Kong, China, 2015, pp. 1929–1933.
[20] D. Priyanga and T. V. P. Sundararajan, “Improved connectivity aware geographical routing protocol in VANETs,”
Journal of Electrical and Electronics Engineering, vol. 2, pp. 1–14, 2015.
[21] C. Bouras, V. Kapoulas, and E. Tsanai, “Performance evaluation of routing mechanisms for VANETs in urban areas,”
Springer International Publishing Switzerland, vol. 2015, pp. 133–153, 2015.
[22] X. Zhang, X. Cao, and Others, “A street-centric opportunistic routing protocol based on link correlation for urban
VANETs,” IEEE Transactions on Mobile Computing, vol. 15, pp. 1586 -1599, 2016.
[23] H. Menouar, M. Lenardi, and F. Filali, “Movement prediction-based routing (MOPR) concept for position-based
routing in vehicular networks,” In Proc. The 66th IEEE Vehicular Technology Conference, Baltimore, USA, 2007,
pp. 2101–2105.
[24] C. Li, L. Wang, and Others, “A link state aware geographic routing protocol for vehicular ad hoc networks,”
EURASIP Journal on Wireless Communications and Networking, vol. 175, pp. 1 –13, 2014.
[25] X. Cai, Y. He, and Others, “LSGO: Link state aware geographic opportunistic routing protocol for VANETs,”
EURASIP Journal on Wireless Communications and Networking, vol. 96, pp. 1 -10, 2014.
[26] Z. Cui, D. Li, and Others, “The next-hop node selection based GPSR in vehicular ad hoc networks,” Journal of
Computer and Communications, vol. 4, pp. 44-56, 2016.
[27] R. Kasana and S. Kumar, “Multimetric next hop vehicle selection for geocasting in vehicular ad-hoc networks,” in
Proc. IEEE International Conference on Computational Intelligence & Communication Technology, Ghaziabad,
India., 2015, pp. 400-405.
[28] C. Woosuk, J. Jongsoo, and Y. Sukdea, “An adaptive link management for vehicular ad hoc networks,” in Proc.
International Conference on Wireless Communications and Signal Processing, WCSP 2010, pp. 1–4.
