An Intelligent Greedy Position-Based Multi-hop Routing Algorithm for Next-Hop Node Selection in VANETs

Abstract

Attempts have been made to discuss how successfully data can be delivered in vehicular ad hoc network which is a special kind of ad hoc network distinguished by its node movement characteristics, hybrid network architectures, and new application scenarios. The research shows that the application of fuzzy logic techniques in multi-hop greedy position-based routing helps in successfully delivering data on roads. A mathematical model has been developed and proposed a fuzzy logic-based greedy routing algorithm (FLGR) for next-hop node selection in vehicular network. FLGR was coded and simulated in MATLAB 7.0 in order to evaluate its performance in terms of average packet delivery ratio and average hop count. The results show that FLGR performs better as compared to other next-hop neighbor node selection methods and helps in delivering data successfully.

Full text

RESEARCH ARTICLE
An Intelligent Greedy Position-Based Multi-hop Routing
Algorithm for Next-Hop Node Selection in VANETs
Shilpy Agrawal
1
•Neeraj Tyagi
1
•Asif Iqbal
2
•Ram Shringar Rao
3
Received: 28 December 2016 / Revised: 9 June 2018 / Accepted: 6 September 2018
ÓIndian Academy of Sciences 2018
Abstract Attempts have been made to discuss how suc-
cessfully data can be delivered in vehicular ad hoc network
which is a special kind of ad hoc network distinguished by
its node movement characteristics, hybrid network archi-
tectures, and new application scenarios. The research
shows that the application of fuzzy logic techniques in
multi-hop greedy position-based routing helps in success-
fully delivering data on roads. A mathematical model has
been developed and proposed a fuzzy logic-based greedy
routing algorithm (FLGR) for next-hop node selection in
vehicular network. FLGR was coded and simulated in
MATLAB 7.0 in order to evaluate its performance in terms
of average packet delivery ratio and average hop count.
The results show that FLGR performs better as compared
to other next-hop neighbor node selection methods and
helps in delivering data successfully.
Keywords Vehicular ad hoc networks 
Intelligent transportation system Fuzzy logic 
Membership functions Position-based routing
1 Introduction
The increase in population density has simultaneously
raised the demand of vehicles for transportation for moving
from one place to another or for any business purpose etc.
This demand of vehicles has lead to the increase in traffic
on roads both in cities and in highways. The increase in the
number of vehicles on roads has posed various challenges
like increase in the number of road accidents, traffic jams,
etc. Also, passengers desire all sorts of safety and enter-
tainment services while traveling. In order to meet these
challenges and to fulfill the passengers requirements,
vehicular ad hoc networks (VANETs) have emerged as a
new network technology. It is an integral part of the
intelligent transportation system (ITS) that aims to improve
a constant demand for information on the current location
and for data, specifically on the surrounding traffic, routes
and much more [1]. In VANETs, vehicles act as mobile
nodes and form a communication ad hoc network. Thus, it
is a subclass of mobile ad hoc network (MANET) where
the nodes are highly mobile in nature because of which
VANET is considered to be a dynamic topology network.
Because of this highly mobile nature of nodes, routing of
data packets from source node to destination node has
become a critical issue in VANETs. Each mobile node
(vehicle) in this network is equipped with a global posi-
tioning system (GPS) device which helps in tracking the
position information of itself and any other vehicles in its
communication range.
VANETs are broadly classified into two types of net-
works, namely vehicle-to-vehicle (V2V) and vehicle-to-
infrastructure (V2I). In V2V, the communication takes
place from vehicles to vehicles, whereas in V2I, any kind
of communication takes place between vehicles and
infrastructure. These vehicular networks have potential to
&Shilpy Agrawal
shilpy15@gmail.com
Neeraj Tyagi
neeraj@mnnit.ac.in
Asif Iqbal
asif@pirotechnologies.com
Ram Shringar Rao
rsrao08@yahoo.in
1
MNNIT-Allahabad, Allahabad, India
2
PIRO Technologies PVT. LTD., New Delhi, Delhi, India
3
Indira Gandhi National Tribal University, Amarkantak, India
123
Proc. Natl. Acad. Sci., India, Sect. A Phys. Sci.
https://doi.org/10.1007/s40010-018-0556-9

enable diverse applications related to safety, entertainment,
and traffic efficiency and management. Some examples of
applications are stationary vehicle warning, overtaking
vehicle warning, regulatory speed limit notification, media
downloading and many more. As routing is one of the big
challenges of these networks, in this paper we suggest a
multi-hop routing protocol in order to address the need for
V2V communications in VANET using short-range V2V
links. The links between vehicles are established through
IEEE 802.11p standard, especially for vehicle-to-vehicle
communication [2].
2 Motivation
The overall goal of our work is to provide efficient and
reliable data delivery in V2V communications. The infor-
mation delivery may involve multiple hops between source
and destination nodes. The source node broadcasts the
message, and all the neighboring nodes help in further
retransmitting the message. This process continues until the
packet reaches destination successfully. But, unfortunately
this may lead to redundant broadcast of data packets in a
dense network. Therefore, we have employed a sender-
oriented protocol approach in which a sender node speci-
fies the next forwarders. V2V communication may lead to
redundant rebroadcast of packets while disseminating
information between vehicles. This broadcast redundancy
may lead to several packet collisions and a higher end-to-
end delay due to medium access control (MAC) layer
contention time. Therefore, focus should be on reducing
this broadcast redundancy by selecting a best neighbor
node to relay a broadcast packet. An inappropriate selec-
tion of relay node leads to packet loss and thus fails in
achieving successful message delivery. Like most of the
existing works [3,4] on efficient relay node selection using
position information and road maps, we too have proposed
an algorithm which selects an approximate best next-hop
neighbor node by employing fuzzy logic. The location
(position) of vehicles is unavailable or imprecise in some
roads such as tunnels, and also sometimes the multiple
metrics used for relay node selection may conflict with
each other which results in uncertainty. Therefore, we have
used fuzzy logic concepts that handle imprecise, uncertain,
inaccurate and incomplete information. Further, we have
also designed a mathematical model that too finds a suit-
able next-hop neighbor node for forwarding the data
packets in order to deliver the packets from source to
destination in vehicular network.
3 Related Work
Multi-hop routing in VANETs demands for efficient han-
dling of next-hop neighbor node selection. In order to meet
this requirement, a small subset of intermediate nodes are
selected to relay data packets and thus reduce the problem
of broadcast redundancy. Therefore, in this section we
briefly describe the fuzzy decision-based routing protocols
used in VANETs that help in selecting such Based on
topology-based protocol ad hoc on demand distance vector
(AODV) [5], a fuzzy control-based AODV routing (Fcar)
protocol has been proposed [6]. Fcar employs fuzzy logic
and fuzzy control methods to make routing decisions under
constrain of two metrics, namely percentage of directional
vehicles and route lifetimes. A novel stability and relia-
bility aware routing (SRR) protocol integrates fuzzy logic
with position-based routing while making packet for-
warding decisions [7] in which distance and direction are
input to fuzzy decision making system so that optimal node
around a smart vehicle is determined for further packet
forwarding.
A relay node selection method considers multiple met-
rics of inter-vehicle distance, node mobility and signal
strength by employing fuzzy logic [8,9]. The protocol
offers a high level of reliability and coherence over existing
options by selecting a subset of appropriate relay nodes. A
fuzzy logic-based multi-hop broadcast protocol for
VANETs (FUZZBR) has also been proposed [4]. In addi-
tion to relay node selection method, a lightweight
retransmission mechanism to retransmit a packet has also
been used when a relay node fails. A fuzzy logic-based
multi-hop broadcast protocol BR-NB (broadcast with
neighbor information) for VANETs has also been sug-
gested [10]. Unlike [4], BR-NB is independent of position
information. BR-NB uses two-hop neighbor information to
infer the inter-vehicle distance and vehicle movement.
In this paper, we propose a fuzzy logic-based multi-
metric next-hop neighbor node selection algorithm. Our
work is different from others as we have employed gaus-
sian membership functions to convert the input numerical
data to corresponding fuzzy value which is suitable for
dynamic networks. Also we have used four node parame-
ters in order to evaluate neighbor nodes. The four metrics
are distance,direction,speed, and position of neighbor
node. Also we have compared our algorithm with our
proposed mathematical model and with the work proposed
in [4] by employing their fuzzy logic concepts in our work
and has given it a name of modified FUZZBR.
S. Agrawal et al.
123

4 Next-Hop Node Selection Using Cost Function
We formulate a mathematical cost function based on
neighbor node’s distance, direction, velocity, and position
from the current forwarding node (CFN) which helps in
deciding and selecting suitable next-hop neighbor node
among several candidate neighbor nodes of source node or
current forwarding node in order to route the data packets
from source to destination in VANETs. In practical prob-
lems, there are more than one parameter which needs to be
optimized in order to get better results. In our work, we
have taken four parameters which need to be optimized so
that jointly together they help in selecting better next-hop
node for forwarding data packets in a multi-hop network.
Jointly considering these four metrics help us in selecting
the better next-hop node which improves routing. A node is
preferable if it is moving with high velocity, making more
progress toward destination (position metric), having less
distance and angle thus following greedy approach. The
cost function is applied on each candidate neighbor nodes,
and whichever neighbor node (NN) has maximum value of
cost function is considered as the next-hop node for further
forwarding the message. The algorithm for next-hop
neighbor node selection using cost function is given in
Fig. 1. The cost function (CFi) is expressed in Eq. (1) as:
CF ¼Velocity þPosition Distance Angle ð1Þ
Therefore, a node with maximum value for cost function
among neighbors of CFN is chosen as next-hop node. The
whole process is repeated until the packet reaches its des-
tination. Based on this mathematical analysis, we have
calculated average packet delivery ratio and average hop
count metrics. However, the disadvantage of cost function
is that either it can be maximized or minimized. Therefore,
it is not certain that the chosen neighbor fulfills the criteria
of being ideal or not. Therefore, we have further proposed a
different technique in Sect. 5for best next-hop node
selection in vehicular ad hoc networks that overcomes the
drawback of next-hop node selection using cost function.
5 Next-Hop Node Selection Using Fuzzy Logic-
Based Greedy Routing (FLGR)
The basic idea of the algorithm has been published
in [11,12]. However, in this paper we have further worked
on a more realistic scenario to evaluate our work and
present our new simulation results. For facilitating V2V
multi-hop communications, wireless radio communication
devices are embedded into all vehicles. We have assumed
that each vehicle in the network is equipped with a GPS
receiver that helps in obtaining the current position of the
vehicle. All the vehicles are moving in the direction of
destination vehicle and also have same transmission range
R. All the vehicles are moving with a varying speed
between 0 and 100 km/h. Also, we have assumed that there
is at least one neighbor node of every vehicle in the net-
work. All the vehicles (nodes) that are within the
VS: Source Node
VD: Destination Node
VCFN : Current Forwarding Node
CF: Cost Function
nNN: Number of Neighbor Nodes of CFN
VBNH: Better Next-Hop Node
SNN: Selected Neighbor Node
1. Let VCFN =V
S
2. Vehicles broadcast HELLO packets in the network
3. VCFN updates its neighbor list
4. If VDis in the transmission range of VCFN
5. VCFN transmits data packet to VD
6. exit
7. for i= 1 to nNN
8. calculate cost function CFi
9. array[i] = CFi
10. endfor
11. Set max= array[1]
12. for i= 1 to nNN
13. if array[i] >=max
14. max= array[i]
15. set SNN =i
16. endif
17. endfor
18. Set VBNH =S
NN
19. VCFN =V
BNH
20. Repeat steps 2 to step 19 till the message is delivered to the destination
Fig. 1 Algorithm for next-hop
neighbor node selection using
cost function
An Intelligent Greedy Position-Based Multi-hop Routing Algorithm for Next-Hop Node Selection...
123

transmission range of a source or current forwarding node
(CFN) are considered as its neighbors. Source/CFN
broadcasts HELLO packets in order to obtain location
information of neighbor nodes. Nodes within the commu-
nication range of source/CFN node receiving HELLO
packet reply back to source/CFN by sending HELLO
packet which contains its position information and other
important information. Thus, in this manner source/CFN
gets aware of the location information of its next one-hop
neighbors.
FLGR instead of broadcasting data packets to all its
neighbors, unicast the packet to one of its selected neighbor
among several neighbors. Thus, the goal of FLGR is to
select an appropriate neighbor node for further forwarding
the packets. The selection of next-hop neighbor node
employs fuzzy logic [13–15]. Typically, any fuzzy logic-
based system involves three steps, namely (1) input, (2)
process and (3) output. The input step is known as fuzzi-
fication. This step converts the input numerical values to
linguistic variables. Linguistic variables are those that take
words as their values rather than numbers. Fuzzification
step uses these predefined linguistic variables and mem-
bership functions (MFs) to convert the input numerical data
to corresponding fuzzy value. The process step maps the
fuzzy values to predefined IF/THEN rules and combines all
the rules to obtain fuzzy output. In fuzzy inference system
(FIS), the IF part and the THEN (or the final) part of the
fuzzy rule are known as antecedent and consequent,
respectively. Finally, the output step converts the fuzzy
output back into a numerical (crisp) value. This is known
as defuzzification. This step uses the predefined output MF
and one of the several defuzzification methods [13]to
convert fuzzy result into a numerical value.
Our work aims to deliver any kind of messages from a
specific source to a specific destination vehicle. If the
destination node is within the transmission range of source
node, then source can directly transmit the information to
the destination. Else if destination is far away from source,
then the packets are forwarded from source to destination
in a multi-hop fashion following greedy approach. Since a
particular CFN can have many candidate neighbor nodes
(NNs), therefore selecting an appropriate next-hop NN for
further forwarding the packets becomes a challenging
issue. Therefore, we have designed a Mamdani FIS that
helps in selecting an approximate ideal next-hop NN. Thus,
the proposed FIS selects an approximate best next-hop
neighbor node out of several candidate NNs. The following
five steps are followed for the selection of an approximate
best next-hop neighbor node:-
1. Determine a set of fuzzy rules for the proposed FIS. In
our work, we have framed 81 rules. Few rules are
defined as shown in Table 1.
2. Fuzzify the four inputs, i.e., distance, direction, speed,
and position metrics using the input MFs. Fuzzification
converts the numerical values of four input metrics
into fuzzy values.
3. The fuzzified values are then input to rule base. The
rule strength for each rule is established by combining
the fuzzified inputs according to the fuzzy rules.
4. For each fuzzy rule, implication method is exploited
for shaping the fuzzy set in the consequent on the basis
of the results of the antecedent. A single number given
by the antecedent serves as an input for the implication
process, and the output is a fuzzy set (truncated output
function). We have used Min–Max method for each
rule in which the minimal value of the IF part is used
as the final degree. Since all the rules are evaluated in
parallel, therefore there is a need to aggregate different
rules. When combining all the rules (aggregation), the
maximal value of the consequents is used.
5. Finally, after obtaining an output distribution by
combining the consequences, we defuzzify the output
distribution to obtain a numerical value as an output.
There are several defuzzification methods like centroid
of area (COA) or mean Of maximum (MOM), fuzzy
mean (FM). We have applied center of gravity (COG)
method to obtain crisp value as an output. This
numerical value gives the score to the neighbor node.
Steps 2–5 are applied to all the candidate NNs in order to
determine their score. The source/CFN then selects the NN
that has maximum value for the output variable to forward
the packet. Thus, the same fuzzy logic-based approach runs
for each CFN in order to select best neighbor as the next-
hop till the packet reaches destination. The algorithm and
flowchart for the same can be referred from [11,12].
6 Metrics for Neighbor Selection
We have considered multiple node metrics in order to
design an efficient algorithm that selects suitable next-hop
node by employing fuzzy reasoning.
(a) Distance Distance metric defines the distance
between a source (S)/CFN and a neighbor
node, say node A, within the transmission
range R(Fig. 2). The Euclidean distance
between the two nodes is given in Eq. (2),
where (x1,y1) and (x2,y2) are location
coordinates of sender and neighbor node,
respectively.
distðAÞ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ðx2x1Þ2þðy2y1Þ2
qð2Þ
S. Agrawal et al.
123

We have classified distance metric into
three ranges, close,intermediate, and far as
shown in Table 2. Distance-based packet
reception probability is considered for the
classification of this metric in order to
improve the packet delivery ratio. The
radio strengths may vary at certain dis-
tances over time in VANETs because of its
highly mobile environment as well as
several interfering objects. We have used
Nakagami-m distribution radio propaga-
tion model [16,17] to describe the radio
wave propagation in vehicular networks in
the absence of any kind of interferences.
Packet reception probability for various
distances is shown in Fig. 3. As can be
seen from Fig. 3, the nodes which are at a
distance of 0–150 m from the sender have
Table 1 Fuzzy-based inference rules
Rules Distance Direction (angle) Speed Position Score
Rule 1 Close Less Low Close Low
Rule 2 Close Less Low Intermediate Medium
Rule 3 Close Less Low Far Medium
Rule 4 Close Less Medium Close Medium
Rule 5 Close Less Medium Intermediate Medium
Rule 6 Close Less Medium Far Medium
Rule 7 Close Less High Close Medium
Rule 8 Close Less High Intermediate Medium
Rule 9 Close Less High Far Medium
Rule 10 Close Medium Low Close Low
Rule 11 Close Medium Low Intermediate Low
Rule 12 Close Medium Low Far Low
Rule 13 Close Medium Medium Close Low
Rule 14 Close Medium Medium Intermediate Low
Rule 15 Close Medium Medium Far Medium
Rule 16 Close Medium High Close Low
Rule 17 Close Medium High Intermediate Medium
Rule 18 Close Medium High Far Medium
Rule 19 Close High Low Close Low
Rule 20 Close High Low Intermediate Low
Fig. 2 Distance between
neighbor and source
Table 2 Linguistic variables for distance metric
Close 0–150 m
Intermediate 150–200 m
Far 200–250 m
An Intelligent Greedy Position-Based Multi-hop Routing Algorithm for Next-Hop Node Selection...
123

maximum of 100% packet reception
probability, but at the same time they are
not near to destination or in other words
much closer to S/CFN. Thus, selecting
them for further forwarding the packets
will lead to much higher hop counts.
Similarly, nodes at a distance between 200
and 250 m are far away from S/CFN (or
close to destination) but have least packet
reception probability (\45%). Thus,
selecting them would lead to greater
chances of packet drop. Due to these
issues, we have given priority to those
candidate neighbor nodes which are at
intermediate distances from the source/
CFN, i.e., nodes at a distance between 150
and 200 m from S/CFN as they have
87–99% packet reception probability and
are comparatively near to destination fol-
lowing greedy approach. Thus, selecting
them will deliver the packets to destination
with less hop count and better packet
delivery ratio.
(b) Direction This metric defines the angle which the
next-hop neighbor node forms between
itself, source/CFN, and destination node.
The minimum the angle neighbor node
forms, the more it has the probability to be
close to the destination node [11,18].
Thus, it follows greedy approach. As
shown in Fig. 4, node Ais selected as a
next-hop node as direction SA is closer to
direction SD than direction SB.
Therefore, \abetween the next-hop node,
current forwarding node and the
destination node can be calculated as
follows [19] as defined in Eq. (3)
cosa¼d2þD2
SD d02
2dDSD
ð3Þ
This input metric too is categorized into
three fuzzy sets as shown in Table 3.
Therefore, a lesser direction factor ensures
that a particular neighbor node is closer to
destination as compared to neighbor that
forms higher angle.
(c) Speed We have considered speed of nodes as
another important input metric for finding
the best next-hop NN. Speed is categorized
as shown in Table 4into three fuzzy sets.
A high speed of the neighbor node is
required to ensure that it reaches the
destination more quickly as compared to
slow moving nodes.
(d) Position This metric helps in determining how
much progress any neighbor node has
made toward destination. It is the ratio of
the distance between source/CFN and the
Fig. 3 Distance versus packet
reception probability for
Transmission Range ¼250
Fig. 4 Direction metric representation
S. Agrawal et al.
123

projection from a next-hop neighbor node
on a straight line joining S/CFN and
destination to the total distance between
S/CFN and destination node [11,19]
(Fig. 5). The position metric is defined as
given in Eq. (4)
PosNN ¼q
Qð4Þ
The higher the value for position metric,
the closer the NN is to the destination, thus
greedy to reach destination. Table 5shows
the linguistic variables for position metric.
The higher value of position metric of
neighbor node ensures that it is closer to
destination.
The current forwarding node uses these MFs of all the
four node metrics of individual neighbor nodes to compute
to which degree these metrics belong to each of the
appropriate fuzzy sets.
7 Results and Performance Analysis
In this section, we have analyzed the performance of our
proposed algorithm FLGR by comparing it with next-hop
node selection method using cost function, next-hop node
selection through fuzzy logic triangular membership
functions [8] and GPSR [20] protocol. For this analysis,
we have considered average packet delivery ratio and
average hop count metrics.
We have presented the simulation work carried out on
MATLAB [21–23] for analyzing the performance of our
proposed techniques in the selection of next-hop neighbor
for routing. Based on the simulation parameters given in
Table 6, we have simulated the algorithm with the variable
number of nodes from 0 to 200. In our simulation, we have
taken randomly one source and one destination node that
are mobile in nature. For the simulation and analysis pur-
pose, city scenario having 1000 400m2road dimension
is considered. The minimum and maximum allowable
vehicle velocity is taken as 0 m/s and 100 m/s. Transmis-
sion range of 250 m is considered for all the vehicles in the
simulation. Nakagami-m radio model parameters are taken
to represent the realistic city network scenario. Table 6
shows the input parameters that we have considered in our
simulation using MATLAB.
7.1 Average Packet Delivery Ratio (PDR) Versus
Number of Nodes
Packet delivery ratio is the measure of total number of data
packets received by destination to the total number of data
packets sent by source over a communication channel.
Figure 6shows that the proposed FLGR algorithm out-
performs other techniques of selecting suitable next-hop
node for forwarding the data packets in vehicular network.
This is because FLGR algorithm evaluates objective
function value of each candidate neighbor nodes of current
forwarding node using gaussian membership functions
Table 3 Linguistic variables for direction metric
Less directed 0–60
Medium directed 60–120
More directed 120–180
Table 4 Linguistic variables for speed metric
Low 0–33 km/h
Medium 33–66 km/h
High 66–100 km/h
Fig. 5 Position metric
calculation
An Intelligent Greedy Position-Based Multi-hop Routing Algorithm for Next-Hop Node Selection...
123

considering inter-vehicle distance, direction, vehicle speed,
and position metrics. FLGR selects the neighbor which
gives maximum objective function value and thus selects
an approximate best next-hop neighbor node for further
forwarding the data packets. Our proposed algorithm
FLGR has better PDR as compared to proposed cost
function, GPSR protocol and modified FUZZBR [4] algo-
rithm for next-hop node selection against the number of
nodes in the network.
7.2 Average Hop Count
Hop count gives the number of nodes through which the
data packets travel between the source and destination
node. Figure 7shows that FLGR algorithm has minimum
hop count compared to GPSR, modified FUZZBR and the
proposed mathematical model. Thus, by our proposed
algorithm the message is going to reach the destination in
less hop count as compared to other three techniques
because our FLGR algorithm selects approximate ideal
next-hop neighbor node at each hop employing fuzzy logic.
8 Conclusion
Message dissemination in VANETs is a crucial task due to
high speed of moving vehicles. Therefore, in order to
overcome with this issue we have proposed FLGR, a fuzzy
logic-based multi-hop greedy position-based routing algo-
rithm for vehicular ad hoc networks. We have also pro-
posed a mathematical model for next-hop node selection.
The performance of FLGR was evaluated through simu-
lation done in MATLAB and compared with that of other
next-hop node selection techniques. From the simulation
results, we concluded that the FLGR was superior to
modified FUZZBR and proposed mathematical model
algorithms in terms of average packet delivery ratio and
average hop count.
References
1. ITSA Website (2012) Official web site of the Intelligent Trans-
portation Society of America. http://www.itsa.org. Accessed May
2012
2. IEEE, IEEE standard 802.11p (2004) Draft amendment: wireless
access in vehicular environments (WAVE), draft1.0
3. Sahoo J, Wu EHK, Sahu PK, Gerla M (2009) BPAB: binary
partition assisted emergency broadcast protocol for vehicular ad
hoc networks. In: Proceedings of 18th international conference on
computer communications and networks, pp 1–6
4. Wu C, Ohzahata S, Kato T (2012) VANET broadcast protocol
based on fuzzy logic and lightweight retransmission mechanism.
IEICE Trans 95–B:415–425
Table 5 Linguistic variables for position metric
Close 0–0.33 m
Intermediate 0.33–0.66 m
Far 0.66–1 m
Table 6 Input parameters
Parameters Values
Simulation area 1000 400 m2
Traffic density 100–200 Vehicles/km
Maximum vehicle speed 100 km/h
Transmission range of vehicle 250 m
Packet size 512 bytes
Radio propagation model Nakagami-m ¼3
Fig. 6 Average delivery ratio versus number of nodes
Fig. 7 Average hop count versus number of nodes
S. Agrawal et al.
123

5. Perkins CE, Royer EM (1999) Ad-hoc on-demand distance vector
routing. In: 2nd IEEE workshop on mobile computing systems
and applications, pp 90–100
6. Wang XB, Yang YL, An JW (2009) Multi-metric routing deci-
sions in vanet. In: 8th IEEE international conference on
dependable, autonomic and secure computing, DASC’09,
pp 551–556
7. Ghafoor KZ, Bakar KA, Salleh S, Lee KC, Mohamad MM,
Kamat M, Arshad MM (2012) Fuzzy logic-assisted geographical
routing over vehicular ad hoc networks. Int J Innov Comput Inf
Control 8(7):5095–5120
8. Wu C, Ohzahata S, Kato T (2010) Fuzzy logic based multi-hop
broadcast for high-density vehicular ad hoc networks. In: IEEE
vehicular networking conference, pp 17–24
9. Wu C, Ohzahata S, Kato T (2012) Fuzzy logic for multi-hop
broadcast in vehicular ad hoc networks. INTECH Open Access
Publisher, London
10. Celimuge WU, Ohzahata S (2013) Practical solution for broad-
casting in VANETs using neighbor information. IEICE Trans
Commun 96(11):2856–2864
11. Agrawal S, Raw RS, Tyagi N, Misra AK (2015) Fuzzy logic
based greedy routing (FLGR) in multi-hop vehicular ad hoc
networks. Indian J Sci Technol 8(30):1–14
12. Agrawal S, Raw RS, Tyagi N (2015) Enhancing greedy routing
using fuzzy logic for vehicular ad hoc networks. In: Proceedings
of the international conference on advances in computing, control
and networking—ACCN, pp 98–102
13. Klir GJ, Yuan B (1995) Fuzzy sets and fuzzy logic–theory and
applications. Prentice Hall, Upper Saddle River (ISBN: 978-0-13-
101171-7)
14. Wang L-X (1997) A course in fuzzy systems and control. Pren-
tice Hall, Upper Saddle River (ISBN: 978-0135408827)
15. Sivanandam SN, Sumathi S, Deepa SN (2007) Introduction to
fuzzy logic using MATLAB, vol 1. Springer, Berlin
16. Taliwal V, Jiang D, Mangold H, Chen C, Sengupta R (2004)
Empirical determination of channel characteristics for DSRC
vehicle-to-vehicle communication. In: Proceedings of the 1st
ACM international workshop on vehicular ad hoc networks
17. Killat M, Hartenstein H (2009) An empirical model for proba-
bility of packet reception in vehicular ad hoc networks. EUR-
ASIP J Wirel Commun Netw. https://doi.org/10.1155/2009/721
301
18. Kranakis E, Singh H, Urrutia J (1999) Compass routing on
geometric networks. In: Proceedings of 11th Canadian confer-
ence on computational geometry, pp 51–54
19. Raw RS, Lobiyal DK (2011) Improved position-based routing in
vehicular ad hoc networks using P-DIR method. In: Abraham A,
Mauri JL, Buford JF, Suzuki J, Thampi SM (eds) ACC (3).
Springer, Berlin, pp 654–663
20. Karp B, Kung HT (2000) GPSR: greedy perimeter stateless
routing for wireless networks
21. Kay MG (2008) Basic concepts in Matlab. http://www.ise.
ncsu.edu/kay/Basic_Concepts_in_Matlab.pdf. Accessed 15 May
2015
22. Nall LO, Hathaway RJ (1996) Fuzzy systems toolbox, fuzzy logic
toolbox [software review]. IEEE Trans Fuzzy Syst 4(1):82–85
23. Uike A, Study and simulation of vehicular ad-hoc network
modules using MATLAB EEL6509 wireless communications
under Dr. Dapeng Oliver Wu
An Intelligent Greedy Position-Based Multi-hop Routing Algorithm for Next-Hop Node Selection...
123