Vehicular ad-Hoc networks (VANETs)—An overview and challenges

Abstract

Vehicular ad-hoc networks (VANETs) technology has emerged as an important research area over the last few years. Being ad-hoc in nature, VANET is a type of networks that is created from the concept of establishing a network of cars for a specific need or situation. VANETs have now been established as reliable networks that vehicles use for communication purpose on highways or urban environments. Along with the benefits, there arise a large number of challenges in VANET such as provisioning of QoS, high connectivity and bandwidth and security to vehicle and individual privacy. This article presents state-of-the-art of VANET and discusses the related issues. Network architecture, signal modeling and propagation mechanism, mobility modeling, routing protocols and network security are discussed in detail. Main findings of this paper are that an efficient and robust VANET is one which satisfies all design parameters such as QoS, minimum latency, low BER and high PDR. Some key research areas and challenges in VANET are presented at the end of the paper.

Full text

Journal of Wireless Networking and Communications 2013, 3(3): 29-38
DOI: 10.5923/j.jwnc.20130303.02
Vehicular Ad-Hoc Networks (VANETs) - An Overview
and Challenges
Sabih ur Re hman*, M. Arif Khan, Tanveer A. Zia, Li h o ng Zhe ng
School of Comp uting and Mathematics, Charles Sturt University, Australia
Abs t ra c t Vehicular ad-hoc networks (VANETs) technology has emerged as an important research area over the last few
years. Being ad-hoc in nature, VANET is a type of networks that is created from the concept of establishing a network of cars
for a specific need or situation. VA NETs have now been established as reliable networ ks that vehicles use for communication
purpose on highways or urban environments. Along with the benefits, there arise a large number of challenges in VANET
such as provisioning of QoS, high connectivity and bandwidth and security to vehicle and individual privacy. This article
presents s tate-of-the-art of VANET and discusses the related issues. Network architecture, signal modeling and propagation
mechanism, mobility modeling, routing protocols and network security are dis cussed in detail. Main findings of this p aper are
that an efficient and robust VANET is one which satisfies all design parameters such as QoS, minimu m latency, low BER and
high PDR. Some key research areas and challenges in VANET are presented at the end of the paper.
Ke y wo rds VA NET, Modeling, Mobility, Routing, Security
1. Introduction
Although Vehicular Ad-hoc Network (VANET) is not a
new topic, it continues to provide new research challenges
and problems. The ma in objective of VANET is to help a
group of vehicles to set up and maintain a communication
network a mong them without using any central base station
or any controller. One of the major applications of VANET
is in the critical medical emergency situations where there is
no infrastructure while it is critical to pass on the information
for saving human lives. However, along with these useful
applications of VANET, emerge new challenges and
problems. Lack of infrastructure in VANET puts additional
responsibilities on vehicles. Every vehicle becomes part of
the network and also manages and controls the
communication on this network along with its own
communication requirements.
Veh ic u la r ad -hoc networks are responsible for the
communication between moving vehicles in a certain
environment. A vehicle can communicate with another
vehicle directly which is called Vehicle to Vehicle (V2V)
communication, or a vehicle can communicate to an
infrastructure such as a Road Side Unit (RSU), known as
Ve hic le -to-Infrastructure (V2I). Figure 1 shows a typical
VANET scenario.
The main contributions of this paper are to present state –
* Corresponding author:
sareh man @cs u. edu. au (Sabih ur Rehm an)
Published online at http://journal.sapub.org/jwnc
Copyright © 2013 Scientific & Academic Publishing. All Rights Reserved
of – the - art in VANET technology. A detailed study of
network architecture with different topologies and network
modeling is presented in this paper. A key design area in
VA N ET i n order to properly form a communication network
is routing the packets in effect ive manner. The paper
discusses different routing algorithms for VANET and
pres e n ts limitations of those algorithms. Security issues in
VA NET environment are also addressed in the paper so that
trustworthy network architecture can be modelled. The paper
also discusses some of the key research areas and challenges
in th is fie ld .
1.1 . Background Research
A very important design aspect of VANET is to develop
an efficient, reliab le and secure routing protocol. Va st
research has been conducted in thisarea[1],[2],[3],[4],[5],
[6],[7],[8],[9],[10],[1 1],[ 12],[13],[ 1 4], [15],[16]. The main
objective of any routing protocol is to find an optimal way of
communication between nodes ( veh icle s ). One popular
VA NET routing protocol is Ad-hoc On Demand Vector
Routing (AODV) p resented in[17, 18] which uses a
demand-driven route estab lis hment procedure. A limitation
of this approach is that it creates flooding type of situation
within the entire network. Many researchers have presented
the remedies for flooding in AODV[19, 20]. Another routing
approach is presented in Dynamic Source Routing[21] which
is also classified as reactive in nature. Routes are stored in
cache and it is expected that source will have co mplete
knowledge of hop-by-hop route to the destination. In[22],
authors have presented another routing approach using bor-
der node concept. This protocol was based on stop and carry

30 Sabih ur Rehman et al. : Vehicu lar Ad -Hoc Networks (VANETs) - An Overview and Challen ges
mechanism.
There are many important applications of VANET in
today’s world. These applications range from critical
medical services to comfort and leisure activities. A VANET
must be able to fulfil all require ment of ever changing user
needs and should also comply with the standard and
architectures of the available technology.
Some of the key applications[9] of VANET can be sum-
marised as follows:
Fi gu re 1. C re at ing an Ad-ho c Net work usin g Veh i cles (VANETs)
Road Tra
ffi
c Safety -Work on reducing the number o f
fatalities/injuries on the roads by alerting the driver about
dangers in advance.
Tra
ffi
c Engineering or E
ffi
ciency - Increase overall
performance of the transport systems by reducing travelling
time and congestion.
Comfort and Quality of Road Travel- Provide comfort
applications for travellers like ‘advanced traveller
informat ion systems’, ‘electronic pay me nt s y s t ems ’,
‘variable mess age signs’ and ‘electronic toll collection’ etc.
Some of the key characteristics of a VANET model are as
follows :
Dyn amic Topology: VANET environment has a
constantly changing topology due to high mobility of the
vehicles. The connection between two vehicles travelling
with average suburban speed limits in opposite directions
lasts for a very short time. This connection time goes much
lesser as the speed of the vehicles increases in a
freeway/highway environment.
Frequent Disconnections: The link connection between
the vehicles in VANET has frequent disconnections because
of the high movement of the nodes and frequent change in
the environmen t .
Mobility Modeling: In order to implement VANET
efficiently and realistically, an accurate mobility model is
required for this highly dynamic environment of VANET.
Predictable Mobility Patterns: In VANET environment
most of the vehicles move on pre-defined roads and
highways. This allows the use of predictable mob il ity
patterns in network design.
Use of Other Technology: Most of the vehicles in VANET
these days are capable of integrating their own system with
other available technologies such as Global Positioning
Sys t em (GP S).
No Power Constraint: As compared to other Mobile
Ad-hoc Network (MANET) devices, the nodes in VANETs
have the privilege of having longer battery life. This can be
utilis ed fo r e fficient processing of comp le x a nd
computational hungry routing/security mechanisms in the
network.
Stringent Delay Constraints: VAN ETs are responsible for
the delivery of critical medical emergency messages. These
messages must be delivered on time in order to save human
lives .
The rest of the paper is organized as follows. Section 2
presents the architecture and network modeling of ad-hoc
ne tworks. Different routing paradig ms and algorith ms along
with their pros and cons are discussed in Section 3. Security
aspects of VANET are discussed in Section 4. Section 5
highlights the key research areas and the challenges in the
field. Section 6 concludes the paper.
2. Architecture and Network Modeling
In princ ipal, there is no fixed architecture or topology that
a VA NET must follow. However, a general VANET consists
of moving vehicles communicating with each other as well
as with some nearby RSU. A VANET is different than a
MANET in the sense that vehicles do not move randomly as
nodes do in MANETs, rather moving vehicles follow some
fixed paths such as urban roads and highways. While it is
easy to consider VANETs as a part of MANETs, it is a lso
important to think of VA NETs as an individual research field,
especially when it comes to designing of network
architecture. In VANET architecture, an on board unit (OBU)
in a vehicle consists of wireless transmitter and receiver.
Fi gu re 2. N et work Arch itecture in VANET
In a broad sens e, we can loosely define three possible
communication scenarios for vehicles. One possibility is that
all vehic les communicate with each other through some RSU.
This architecture may resemble wireless local area networks
(WLAN). Second possibility is where vehicles directly
communicate with each other and there is no need of any
RSU. This can be classified as Ad-hoc architecture. In third
possibility, some of the vehicles can communicate with each
other directly while others may need some RSU to
communicate. Th is can be referred as hybrid scenario[23].
Figure 2 shows these three possibilities.
Understanding of network architecture is important in
order to realise the full potential of vehicular communication.
Most of the researchers[24, 25 ] hav e based their studies by
dividing VANET s cenarios in three categories namely Urban,

Journal of Wireless Networking and Communications 2013, 3(3): 29-38 31
Rural and Freeway/Highway.
One of the reasons to investigate in such a manner is to
make sure that it will eventually cover the need of
inter-networking for entire vehicular environment. Each
environment has its own specific challenges to overcome.
For example, in a sparse network like highways, the low den-
sity of vehicles remains the prime issue. Even in some urban
environments, low penetration ratio and low traffic at night
times can cause long network delays.
One of the most important attributes of mobile ad-hoc
wireless network is the mobility associated with the nodes.
The highly mobile nature of vehicles makes it very
complicated to model the communication scenario. Mobility
model for a VA NET environment has to look deeper into key
characteristics of vehicular mobility like acceleration,
deceleration, changing lanes and human driving patterns. For
this purpose, an ample amount of research[26, 27, 28 ] h a s
been made to include mobility in the design of VANET. In
VA NET paradigm, the mobility model must include the
behaviour of moving vehicles individually and in a group for
an erro r free e fficient packet transmission.
Generally, a mobility model can be classified at micro and
macro levels depending on the nature of details required in
modeling. The consideration for roads, buildings and streets
are classified as macroscopic whereas the behaviour of
vehicles is treated as mic roscopic level[9]. In general, the
mobility models in VANET can be categorised in three
categories such as: a) Stochastic or Random Modeling, b)
Traffic or Flow based Modeling, and c) Trace based Mod-
elin g [28]. Th is ranges from the fluid flow theory to stochastic
theory which includes random walks and the aggregated
behaviour of the vehicles. On the other hand, the mobility
modeling can also be based on the degrees of randomness in
the model one wish to include. Many researchers used
deterministic approach, hybrid approach or total random
approach to model the dynamic nature of vehicles in a
VA NET. Mobility modeling impacts on the performance
an a lys is of the networks such as on end-to- end delay,
capacity, security, routing, and scalability of the networks
few to name[2 9].
A more concrete network modeling for ad-hoc networks
can be done using appropriate mathematical and statistical
tools[30]. In one of the approaches, a wireless ad-hoc
network can be modelled as a random graph with vertices of
the graph representing nodes and edges in the graph may
represent commun ication links between the nodes[31]. A
graph model of mutli-hop wireless network can be
represented as G = (V, L) where V is the vertex set and L is
the edge set of the graph. Such a graph can be constructed
using Random Geo metric Graph (RGG), where vertices are
independent and identically distributed in a specific area
satisfying the Poisson Distrib u t io n . Th is mod el is s imilar to
the connected graph of a Boolean Model (BM)[32]. A mo re
theoretical and analytical understanding of this topic has yet
to be done for ad-hoc wireless netwo rks .
One of the most challenging areas of research for VANET
is the actual transmission and reception of the signal. The
wireless channel through which the signal propagates is
highly dynamic in nature. It is full of different scatterers,
reflectors and absorbed by many objects during the
transmission. Therefore, it is of utmost importance to
properly model the signal which helps in correct recovery of
the transmitted signal. At the receiver side, generally signal
does not take single path to arrive. There are mult i-paths for
the signal to arrive at the receiver. One of the challenges is to
combine these different components of the received signal in
such a way that the correct data must be recovered.
Let us consider that  ∈   is the complex Gaussian
received signal with zero mean and unit variance at one of
the nodes and is given by the following equation
= + , (1)
where  ∈ × is the co mp lex channel gain matrix,
 ∈  is the co mplex trans mitted signal and  ∈   is
additive white Gaussian noise. In order to minimise the bit
error rate and provide the promised QoS, one has to reduce
the effect of wire less channel on the received signal. For this
purpose a number of wireless channel models such as Free
Space Model[59], Nakagami Model[58] and Log Normal
Shadowing Model have been considered. These models also
play an important role in creating the simu lation
environment for different VANET scenarios. For example,
NS-2 uses Nakagami model whereas OMNET++ by defau lt
uses Free Space Model. In any VANET environ ment,
calculating the transmission range and then finding the
probability of packet delivery ratio (PDR) is important.
Transmission range for a Free Space wireless channel model
can be calculated using Friis Transmission equation[60]
given as

=   2
(4  )2 
 , (2)
where 
 and 
 are receive and transmit powers
respectively,  and  are gains of transmit and receive
antennas respectively, λ is the wavelength and  is t h e
distance between transmitting and receiving nodes.
Considering proper propagation scheme, wireless channel
model and combining schemes at the receiver imp roves the
QoS in VA NET.
3. Routing in VANET
One of the major challenges in the design of veh ic u lar
ad-hoc network is the development of a dynamic routing
protocol that can help disseminate the information from one
node (vehicle) to another. Routing in VA NET is different to
the traditional MANET routing because of highly dynamic
and ever changing topologies in the former. Few protocols
that were earlier designed for MANET environment have
been tested on VANET. The challenge however remains as
how to reduce delay associated with passing the information
from one node to another. Overcoming these hurdles in
MANET protocols, can help implement real time appli-
cations for VANET environment. Other implications such as
reducing control overheads also need to be looked into

32 Sabih ur Rehman et al. : Vehicu lar Ad -Hoc Networks (VANETs) - An Overview and Challen ges
carefully. Keeping an eye on the dynamic characteristics of
VA NET (as highlighted previously), the routing protocol
should be able to withstand the unpredicted and dynamic
nature of vehicular network topology. Perhaps the most
difficult task in VA NET routing is finding and maintaining
the optimal paths of communication in desired environments
[6].
Most of the routing protocols in VA NET are closely
lin ked wit h the topology being used in the network
architecture and the performance deviates whenever there is
a change in network topology.
As highlighted in Figure 3, routing in VANET can be
classified into five major categories namely as:
• Ad-hoc or Topology Driven Protocols
• Location Based Routing Protocols
• Cluster Based Protocols
• Broadcast Protocols
• Geocast Protocols
3.1 . Ad-hoc or Topol ogy Dr i ve n Routi ng
In general, VA NETs are inf rastructure -less networks a n d
many routing protocols devised for prior ad-hoc network
such as MANET based on different network topologies may
be applied to VA NETs with certain modifications. Topology
driven protocols are sub-classified into three categories such
as proactive, reactive and hybrid. A number of such
protocols were designed to cater the needs of VANET
environmen t [17],[2 1] ,[22], [33],[34]. In a proactive protocol,
nodes continuously update their routing table with
information regarding ne w routes within the network. This
informat ion is passed around to all nodes by sending periodic
HELLO pac kets. This approach, however, creates su bs tantial
control overheads. This restricts the use of limited wireless
resource such as available bandwidth.
On the other hand, in reactive approaches, for example
AODV[17], DS R[21] , BRP [22] nodes will only send the
control data when there is a need. This reduces overheads
associated with establishing the link, and helps distribute the
actual information faster. This approach however still puts
undue resource overheads like maintenance of used/un us ed
routes. These unused paths are created and broken, due to
s t ringent network topology of VA NET. Overheads created in
reactive protocols are associated with discovering the path to
send the information. The path finding process is initiated by
sending certain type of message called Route Request
Message (RREQ).
In Figure 4, node (S) wants to send information to node
(D). This process will be initiated first by discovering of
route to the destination. When node S needs to find a route to
node D, it broadcas ts an RREQ mess age to all its neighbours.
When intermediate nodes, say node 1 receive a RREQ
message, they compare the desired destination with their
own identifiers. If there is a match, it means that the request
is destined for node 1, otherwise, node 1 will rebroadcast the
RREQ to its neighbours and so do all the other nodes. This
approach can create a flooding in the network. Once the
request reaches the destination (node D in this case), Route
Reply Mess age (RREP) is initiated back to the source using
Backward learning method.
Besides studying basic reactive and proactive type of
protocols, researchers have found tremendous liking to
discover hybrid protocols as well. An examp le of such a
protocol is discussed in detail in[33]. In this approach, the
authors have focused more on the design architecture of
whole network rather than the performance analysis.
Fi gu re 3. Ro utin g Types in VANET

Journal of Wireless Networking and Communications 2013, 3(3): 29-38 33
Fi gu re 4. Traditional Rout ing -Route Discovery using Flooding
One protocol that has shown robust behaviour against the
topology change is Temporally Ordered Routing Algorithm
(TORA)[35]. It sends localised routing messages to a small
set of node in the vicinity of the change. In TORA there is a
potential for oscillation to occur, especially when various
sets of nodes are trying to be a part of one space set.
3.2 . Loc atio n B ased R outing
Another category of protocols that have shown interest
among the researchers are Location or Position Based
Routing protocols. In this scheme of protocols information
regarding geographic location of vehicles is obtained from
different sources like maps, Global Positioning System
(GPRS) or even traces of traffic models to help disseminate
the information. Quite a few researches like[36] and[37]
have presented a thorough comparison of well known
topology based protocols like AODV and DSR in con-
junction with Position Based algorithm and the results have
shown better and improved performance as compared to
using plain topological approach.
Unlike topology based protocols, position based protocols
do not need any route maintenance and the route can get
established when there is a need for it. This reduces undue
constraint on the bandwidth that is already inadequate in
VA NET environment. Greedy Perimeter Stateless Routing
(GPSR)[38] is one of the examples of a location based
protocols where search for nearest neighbour is performed.
Each node is aware with its position as well as the position of
its clos est neighbours. Having the main node this specific
information makes GPSR highly suitable for dynamic
topological networks. In scenario where no nearest neigh-
bour exists, GPSR utilises perimeter mode with face routing
to maximise the search for destination. A performance
comparison of GPSR and DSR in a highway environment
has been presented in[37]. It has been shown that an
improvement in route delay is achieved while using GPSR.
In applications where delay and Packet Delivery Ratio
(PDR) have stringent requirements mostly in urban
environment, GPSR in its original form cannot be used. To
cater this problem,[39] presented a modified version of
GPSR called Advanced Greedy Forwarding (GPSR-AGF).
In this version, information about speed and direction of the
destination node are included within the HELLO packets.
This showed an improvemen t in re lation to the above
mentioned performance metrics. To overcome the
deficiencies of GPSR, another approach was presented by
the introduction of Geographic Source Routing (GSR)[40].
In GSR, road layouts are used to discover the destination
route. It mostly relies on availability of road maps which are
used in conjunction with the shortest path finding algorit h ms .
Another important location based routing approach was
presented in[12],[13] namely Location Assisted Routing
( LA R). LA R is cla s sified as a reactive protocol which us es
location information of nodes to decrease the routing
overheads that were underlined by other reactive protocols
such as AODV and DSR. LAR mainly employs two methods
to determine the next hop, one based on window size and
other with distance variation.
In[13], authors have presented a study of LAR scheme for
VA NET in Highway scenario using the performance metrics
s u ch as Packet Delivery Ratio (PDR), System Throughput,
Average Delay and Routing Overheads. It is shown that the
PD R in creases for moderate number of nodes but decreas es
at a higher network density. The reason being that at higher
network density, the connectivity between nodes is high
resulting in less loss of packets. The system throughput of
LA R increases from low to moderate network density but it
decreases when the number of nodes is higher than 50 in the
network. The reason for decrease in system throughput could
be the high in te rfe rence among the transmitted signals .
Routing overheads and average delay decrease with the large
number of nodes because of many routing paths and larger
number of hops.
A different methodology not relying on external resources
like maps or infrastructure is presented in Greedy Perimeter
Coordinator Routing (GPCR)[41]. GPCR is based on the u-
derstanding that street paths in urban environment create a
planner graph. Packets are always forwarded to a node
(coordinator) that is on the junction edge. The routing deci-
sions are only made by the coordinator. In another routing
scheme where routing decisions are made based on the
connected paths between source and destination pair is
Connectivity Aware Routing (CAR)[42]. Anchor-Based and
Traffic Aware Routing (A-STA R) [36] is another position
based protocol used for urban environments. It exploits the
usage information for a particular route, to determine the
anchor path (number of junctions) to establish maximu m
connectivity. This usage can be determined us in g s t atist ic a l
data gathered for particular surroundings. If packet
(information) is needed to be sent to any node in between an-
chors, the usual greedy approach is used. Studies have shown
a much improved performance by using A-STA R as
compared to GSR and GPSR[43].
3.3 . Clus ter Base d Routing
In order to reduce the network traffic and routing

34 Sabih ur Rehman et al. : Vehicu lar Ad -Hoc Networks (VANETs) - An Overview and Challen ges
overheads in VANET, a routing paradigm namely Cluster
Based Routing (CBR) is introduced in[44] ,[4 5], [46],[47].
The main idea behind CBR is to create a network
architecture based on s mall groups of vehicles called as
clusters. In a cluster, one of the vehicles plays the role of a
cluster-head as shown in Figure 5. The size of the cluster
depends on the design of the routing a lgorith m wh ich may be
based on the number of vehicles in a cluster or the
geographical position of the veh icles .
Fi gu re 5. Cl ust er Based Routin g
The concept of CBR is not new to ad-hoc networks. There
are studies in literature to test the performance of CBR in
MANETs where network conditions are rigid[44]. Most of
these routing protocols are based on first establishing a
mechanis m or protocol to create the cluster range, and then
looking into searching the optima l routes for communication.
Due to scalability issue, the cluster formation techniques in
MANET can’t be directly applied to VANET.
In LO RA -C BF [45], authors have presented a reactive
geographic routing based algorithm for V2V communication
by introducing one cluster head, zero or mo re me mb e r s o f
the cluster and one or more gateway to communicate with
other cluster-heads. The architecture of the LORA-CBF
helps to minimize the unnecssary overheads of
retransmissions. Firstly, as in reactive approach, routes in
LO RA -CBF are updated only when there is a need for it to be
updated. Secondly, it uses a concept of having vehicular
gateways to pass control information and hence reducing the
control traffic overheads. Cluster-head is responsible for its
cluster and sends periodic beaconing messages to update its
parameters. Only gateways and cluster-heads are allowed to
retransmit the route discovery packets, Location REQuest
(LREQ) packets where LREQ are packets same like RREQ
packets in topology driven protocols such as AODV and are
transmitted to gather location information of other clusters.
LO RA -CBF can be suitable for highly mobile environment
li ke VA N ET b ecause it updates the location information of
clu s ters rather than the individual vehicles. Simulation
results have shown better performance by LORA-CBF as
compared to other reactive routing protocols like AODV and
DSR[45] .
In[4 6] , a u thors have presented another approach in
designing of a cluster based routing protocol where they
divided the included geographic area into grids. Depending
on number of vehicles in the grid area, the vehicles are
selected as leader of the group. When the source vehicle
wants to send information to destination, it doesn’t need to
discover the whole route, it just forwards the data to optimal
neighbour cluster-head. The data will then be disseminated
to the destination (in the cluster) by the cluster head. This
minimises the overheads associated with the data dissem-
ination and saves the memory space needed to store routing
information.
In[47] authors have presented another cluster based
routing scheme with the main objective of achieving better
results for network topology stability as well as decreasing
the dynamic natu re of VAN ET. The routing algorithm
co ns id ers the speed differences among the vehicles in
addition to the geographical location and destination of the
vehicles. The performance of proposed threshold based
algorithm is compared with other position based and weight
based algorithm in terms of the change of average number of
clusters. It is shown that the average cluster change per
vehicle with the proposed algorith m is much smaller
compared to other routing protocols. This concludes the less
dynamic nature of the algorithm.
3.4 . Broa dc as t Routi ng
Broadcast Routing was one of the traditional routing
techniques used in VANET. Primarily broadcast approach is
used when the message is needed to be sent to the vehicle
that is outside the range. Packets are transmitted using
flooding techniques. This ensures delivery of information,
but uses extensive resources of bandwidth. As briefed
previously, this sort of technique is utilised in many well
established routing protocols, especially in the stage of
discovering of route to the destination. BROADCOMM[48]
and the Nth-Powered P-persistent Broadcast protocol (NPPB)
[49] are such well known protocols designed using the
broadcasting concept.
In BROADCOMM, a hierarchical structure of highway is
simulated and the whole region is divided into virtual cells.
Ce ll me mb e rs es tab lish a hierarchy of Cell Reflector ( CR),
which acts like a base station to gather mess ages for
particular call as well as from neighbouring cells. CR helps
in making judgements in relation to forwarding the messages
to individual vehicles. This protocol has shown an imp roved
performance in message broadcasting, delay and routing
overheads when compared with other broadcasting protocols.
In N PP B[4 9] a probabilistic broadcasting approach is
designed to mitigate broadcasting storms in dense VANET
in order to transmit emergency messages effectively. The
authors studied a weighted p-persistent routing scheme
which shows better accessibility of the farthest node. The
performance of this protocol depends on a reasonable choice
of the number of nodes in the environment.
3.5 . Geoc as t Routing
Geocast routing is the classification of routing that deals
with d issemination of information in specific area of
relevance. Since the early induction of VANET, quite a few
approaches of Geocast routing werepresented[10],[11 ],

Journal of Wireless Networking and Communications 2013, 3(3): 29-38 35
[12] ,[13],[14],[15], [16]. Many VANET applications require
position dependent multi cas ting e.g . d iss emin at inghazar
dous traffic information to vehicles approaching in the same
direction. The key idea behind the Geocast routing is to
narrow down the search for next hop to a specific Zone Of
Relevance (ZOR). Imagine the possibility of having a
mechanism in which if a car gets involved in an accident, it
will automatically report the accident to the approaching
vehicles within that zone.
In[10], authors have presented a Geocast routing protocol
by submitting design architecture of RObust VEhicular
Routing (ROVER). In ROVER, only control packets are
flooded in the network and the data is disseminated using
unicast approach that in turn increases the efficiency and
reliability of the routing scheme. In ROVER, a ZOR is
created as a rectangular window at the back of the source
node. The dimensions of the window created are defined as
length (L) and width (W) respectively. The size of the
window is decided such that all of the lanes on the current
road are covered. The node (vehicle) will accept the mess age
if it rece iv es it wh ile it was in ZOR and re ject it if it is no t
within that zone. It was shown in[10] that ROVER presents
better results for low density environments.
In[14], authors have presented a geographic routing
scheme, GROOV, for both city and highway scenarios based
on the geographical routing strategy presented in[15].
GROOV is based on the idea of finding the most suitable
relay node rather than on the greedy selection criteria. Most
of the greedy selection schemes a re based on improving the
average end-to-end delay while compromising on PDR. It is
shown that GROOV achieves higher PDR co mpared to the
routing algorithm presented in[15] for city environ ment.
In order to localize the search space by using node
positions and map information, there are few techniques
which a re based on this idea. In[16], authors have presented a
context based routing protocol, VCARP, which uses vehicles
information such as destination, location and packet cache
state to make optimal routing decision. The packet
retransmissions are avoided by storing them temporarily in a
cache. Simu lations studies have shown that VCARP
achieves better PDR and reduces the routing overheads in
VANET scenarios. Another vehicle assisted routing protocol
for VANET is presented in[50] and is ca lled Vehicle Assisted
Data Delivery in Ve h icu lar Ad -hoc Network (VADD). In
this protocol authors adopted the concept of carry and
forward. Four different variants of VADD are presented and
their performance is compared with each other. It is shown
that the proposed VADD protocols have better performance
in terms of PDR, Average End to End delay and routing
overheads compared to other existing protocols. Among the
different variants of VADD, the H-VADD out performs the
other three.
4. Security in VANET
Security in VANET should be considered as important as
securing other networks in computing. Due to the highly
sensitive nature of information be ing broadcasted via
VA NET, all applications designed for vehicular network
need to be protected from malicious man ipulation. Imagine
the possibility of a critical message been manipulated and the
harm it will cause if not detected. In addition to that, comfort
and quality applications in VANET need to be protected to
prevent loss of revenue. Va rious consortiums like Network
on Wheels ( N OW ) [51] and Secure VehicularCommunicat
ions (SEVECOM) [52] are running with prime focus of
addressing security issues within the vehicular networks.
As per basic computer and network security definit ions
[53], attacks on a computer network can be classified in three
main groups of threats; threats associated with Authenticity,
Confidentiality and Availability of the resource. If one
applies this model of security on vehicula r network, the one
threat that really stands out is the Confidentiality of the
source.
For example, an attacker who is busy in analysing which
certificates are attached to each message distributed in the
system, might also be able to track the exact location of the
vehicle (compromise of privacy). Currently a broadcast
authentication scheme is utilised in current standards of
security for VA NET such as IEEE 1609.2. This scheme is
based on the use of a public key signature. Broadcast
authentication enables the receivers to verify that received
informat ion was really s ent by the cla imed sender. In order to
protect the privacy of the node, few approaches have been
adopted[54] ,[55].
In[54], authors presented an approach of loading each
vehicle w ith mult iple ce rtified p ublic keys instead of only
using one default key. In another approach[55], authors have
presented a scheme of using pseudonym pools of identifie rs
in which vehicles are allowed to s witch between different
identities. One limitation of these approaches is that they put
undue processing constraints on the network that is already
being pushed to the limit due to the lack of bandwidth. In
order to understand how to develop a secure vehicular
network, one has to really consider the ‘Attacker Model’ and
the types of attacks which are encountered in VANET.
Usually attackers are led by their own motives like curios it y ,
malignity or just a competitive spirit. An attacker might have
access to inside knowledge and can be class ified as an inside
attacker which poses same level of threat as an outside
attacker.
A better approach in VANET security could be to provide
an authentication mechanism to each node. In[56], authors
have presented such a security method which encourages the
nodes to provide a secure sender authentication. Due to the
large number of independent network members (vehic les)
and the existence of human factor, misbehaviour can take
place. So an authentication trust needs to be established.
In VANET security, the attack threats can be classified
into different categories. In[57], authors have described three
key types of attacks:
Bogus Information:- An inside attacker can make bogus

36 Sabih ur Rehman et al. : Vehicu lar Ad -Hoc Networks (VANETs) - An Overview and Challen ges
safety messages to be distributed in the entire network. This
can cause disastrous situations (a threat to Authenticity).
ID Disclosure:- Location information in relation to
vehicle’s exact position (privacy) needs to be protected (a
threat to Confide n t ia lit y ).
Denial of Service:- Attackers can potentially flood the
entire network so that no one will be able to use the applica-
tions/services. Such circums tances can create catastrophic
s it u atio ns if tr igg e re d in s tantaneously (a threat to
Availab ility ).
The two key challenges in re lation to providing a secure
communication in VA NET can be briefly classified as
establishing a robust s ys tem of sender authentication and
providing a mechanis m to keep the user location undisclosed.
Since vehicles are boosted with sufficient processing power,
the computational resources needed for applying
cryptographic techniques in real-test bed should not be a
concern. On the other hand, if needed to be implemented in
virtual (simulated) environment, computational resources
required such as speed of processor and desired memory will
need to be looked into.
5. Ke y Research Areas and Challenges
Although there has been an ample amount of research in
VANET, still there are many areas which need to be looked
into. Due to the different nature of VANET form many other
wireless communication networks and hard design re-
quirements, there are many interesting research problems in
this field. The paper s ummarizes some of the key research
areas and challenges as under. However, it must be noted that
the research challenges in VANET are not limited to only
these areas.
(a) Quality of Service (QoS): Provision of certain quality
of service levels in VANET is an important task. A network
with min i mum delay for data delivery, less retransmissions,
and high connectivity time can provide certain QoS
guaranteed to the users. Promising this kind of QoS with
different user applications and dynamic network en-
vironment is an interesting and challenging task in VANET
design.
(b) E
ffi
cient Routing Algorithms Design: In order to
timely and properly sending data packets from one node to
another node an efficient routing algorithm is required. In
VA N ET , e fficient routing algorith m means a routing scheme
with min imu m delay, ma ximu m system capacity and less
computation al comp lexity. Design such an algorithm which
can be implemented in multiple topologies of the network
and satisfies all of the above mentioned properties is an
active area of research in VA NET.
(c) Scalability and Robustness: Designing a scalable and
robust network remains an open area of research in VANET
because of its challenging characteristics. Many design
approaches fall short when VANETs transform fro m spars e
to high dense mode, or from high mobility to slow traffic
scenarios. A complete VANET framework that is scalable to
different network scales and robust to the topological
changes is required. This is an emerging area of research for
VA NET environment.
(d) Co-operative Communication: A key challenge in
VAN ET is es tab lishing the communication among different
nodes. Different concepts of co-operative communication
fro m wire less network theory may not be directly applied to
VA N ET . T h is c o -operative commun ication, such as up to
which extent nodes should e xchange information among
themselves, is one of the key research areas in the VANET
design.
(e) Network Security: As the nodes in VANET
environment seek exchange of information among each other
all the time, making sure that certain c r itica l p riv acy
information rema in s wit h in the concerned node is an im-
portant design aspect. Designing a proper authentication
mechanism and a trust based security protocol is an
interesting and open research area in VANET.
6. Conclusions
This paper presented an overview and tutorial of various
issues in VANET. Various types of research challenges are
highlighted in context of vehicular communication. In
particular, this paper presented a review of VANET
architecture, transmission modelling, mathe matical aspects
of signal modelling, routing protocols and security. A
comparative analysis of different routing algorithms in the
field of VANET has been presented. It also highlighted the
main issues in routing algorithms. The performance metrics
for routing algorithms, discussed in this paper, were PDR
with respect to average velocity of vehicles, node density and
system throughput. The other parameters of interest
discussed widely in the paper were average end-to-end delay
and routing overheads. The paper concluded that some
algorithm perform we ll in urban environment while others
are suitable for highway environment. It was also concluded
that proper modeling techniques are necessary for designing
a seamless communication in VANET for a particular
environment. Finally, main research challenges and areas of
interest in vehicular commun ication were discussed.
REFERENCES
[1] K. C. Lee, U. Lee, and M. Gerla, “Survey of Routing Proto-
cols in Vehicul ar A d H oc N et works,” in Advances in
Vehicular Ad-Hoc Networks: Developments and Challenges,
M. Watfa. IGI Global, 2010, ch. 8, pp . 149–170.
[2] C. Sommer and F. Dressler, “The dy mo routing protocol in
vanet scenarios,” in Vehicular Technology Conference, 2007.
VTC-2007 Fall. 2007 IEEE 66th, 2007, pp. 16–20.
[3] T.-Y. Wu, Y.-B. Wang, and W.-T. Lee, “Mixing greedy and
predictive approaches to improve geographic routing for
vanet,” Wireless Communications and M obile Comp uting,
vol. 12, no. 4, pp. 367–378, 2012.

Journal of Wireless Networking and Communications 2013, 3(3): 29-38 37
[4] J. Bernsen and D. M anivannan, “Unicast rout ing protocols for
vehicular ad hoc networks: A critical comparison and classi-
fication,” Pervasive and Mobile Computing, vol. 5, no. 1, pp.
1 – 18, 2009.
[5] R. Kumar and M. Dave, “A comparative study of various
routing protocols in vanet,” CoRR, vol. abs/1108.2094, 2011.
[6] F. Li and W. Yu, “Routing in Vehicular Ad Hoc Networks: A
Survey,” Vehicular T echnology M agazine, IEEE, vol. 2, no. 2,
pp. 12–22, 2007.
[7] S. Jaap, M. Bechler, and L. Wolf, “Evaluation of routing
protocols for vehicular ad hoc networks in city traffic scenar-
ios,” in in Proc of the 11th EUNICE Open European Summer
School on Networked Ap p lications, Colmenarejo, 2005, pp .
584–602.
[8] J. Broch, D. a. Maltz, D. B. Johnson, Y.-C. Hu, and J.
Jet cheva, “ A performance comparison of multi-hop wireles s
ad hoc network routing protocols,” Proceedings of the 4th
annual ACM/IEEE international conference on M obile
computing and networking -MobiCom ’98, pp . 85–97, 1998.
[9] H. H art enstein and K . Laberteau x, VANET Vehicu lar A p -
plications and Inter-Net working T echno logies, ser. Int elli gent
Transport Systems. Wiley, 2009.
[10] M. Kihl, M . L. Sichitiu, and H. P. Joshi, “Design and Eval-
uation of two Geocast Protocols for Vehicular Ad-hoc N et -
works,” Swedish Governmental Agency for Innovation Sys-
tems (Vinnova), 2007.
[11] H. Rahbar, K. Naik, and A. Nayak, “DTSG: Dynamic
time-stable geocast routing in vehicular ad hoc networks,” in
Ad Hoc Networking Workshop (M ed-Hoc-Net), 2010 The
9th IFIP Annual M editerranean, 2010, pp. 1–7.
[12] Y. B. Ko and N. H. Vaidya, “Location-Aided Routing (LAR)
in mobile ad hoc networks ,” ACM journal of Wireless Net-
works, vol. 6, pp. 307–321, 2000.
[13] A. D. Akhtar Husain, Brajesh Kumar, “A study of location
aided routing (LAR) p rotocol for vehicular ad hoc networks
in highway scenario,” International Journal of Information
Technology and Web Engineering, vol. 2, no. 2, pp. 118–124,
2010.
[14] S. K. Dhurandher, M. S. Obaidat, D. Bhardwaj, and A. Garg,
“GROOV: A geographic routing over vanets and its p erfor-
mance evaluation,” in Global Communications Conference
(GLOBECOM), 2012 IEEE, 2012, pp. 1670–1675.
[15] Yan-Bo Wan g; Tin-Yu Wu ; W ei-T song Lee; Chih-Hen g Ke,
"A Novel Geographic Routing Strategy over VANET,"
Advanced Information Networking and Applications
Workshops (WAINA), 2010 IEEE 24th International
Conference on , vol., no., pp .873,879, 20-23 April 2010
[16] R. B. Soares and A. A. F. Loureiro, “VCARP: Vehicular
ad-hoc networks context-aware routing protocol,” 2012 IEEE
Symposium on Comp ut ers and Communications (ISCC), vol.
0, pp. 000 442–000 447, 2012.
[17] C. E. Perkins and E. M . Royer, “Ad-hoc On-Demand D is-
tance Vector Routing,” in Second IEEE Workshop on Mobile
Computing Syst ems and Applications, 1999, pp . 90–100.
[18] J. H¨arri, F. Filali, and C. Bonn et, “ Performance comparison
of AODV and OLSR in VANETs urban environments under
realistic mobil it y patterns ,” in M ed-Hoc-Net 2006, 5t h
Annual Mediterranean Ad Hoc Networking Workshop, S.
Basagni, A. Cap one, L. Fratta, and G. Morabito, Eds. IFIP,
Jun. 2006.
[19] B. Li, Y. Liu, and G. Chu, “Improved aodv routing protoc o l
for vehicular ad hoc networks,” in Advanced Computer
Theory and Engineering (ICACTE), 2010 3rd International
Conference on, vol. 4, 2010, pp . V4–337–V4–340.
[20] S. Eichler, F. Dtzer, C. Schwingenschlgl, F. Javier, F. Caro,
and J. Eberspcher, “Secure routin g in a vehicular ad hoc net -
work,” in in Proceedings of the 2004 IEEE 60th Vehicular
Technology Conference, 2004.
[21] D. B. Johnson and D. A. Maltz, “Dynamic Source Routing in
Ad Hoc Wireless Networks,” in Mobile Comput ing, 1996, ch.
5.
[22] M. Zhang and R. S. Wolff, “A Border Node Based Rout ing
Protocol for Partially Connected Vehicular Ad Hoc Net-
works,” Journal of Communications, vol. 5, no. 2, pp. 130–
143, Feb. 2010.
[23] M . Wat fa, Advances in Vehicular Ad-Hoc Networks: Devel-
opments and Challenges, ser. Intelligent Transport Systems.
IGI Global, 2010.
[24] C. Lochert, H. Hartenstein, J. Tian, H. Fussler, D. Hermann,
and M. Mauve, “A routing strategy for vehicular ad hoc net-
works in city environments,” in Intelligent Vehicles Sympo-
sium, 2003. Proceedings. IEEE, 2003, pp. 156–161.
[25] A. Fonseca and T. Vazo, “Applicability of position-based
routing for {VANET} in highways and urban environment,”
Journal of Network and Comp uter Applications, vol. 36, no. 3,
pp. 961 – 973, 2013.
[26] F. Sommer, C & Dressler, “Progressin g toward Realistic
Mobility M odels in VAN ET Simulat ions,” IEEE
Communications Magazine, no. November, pp . 132–137,
2008.
[27] J. Harri, F. Filali, and C. Bonn et, “M obility models for
vehicular ad hoc networks: a survey and taxonomy,” IEEE
Communications Surveys & Tutorials, vol. 11, no. 4, pp .
19–41, 2009.
[28] M . Fiore, P. T orino, H. Jerome, F. Filali, C. Bonnet, and I. Eur,
“Vehicular Mobility Simulation for VANETs Department of
Mobile Communications,” Simulation, 2007.
[29] F. Baccelli, B. Blaszczyszyn, and P. Muhlethaler, “Stochastic
analysis of spatial and opportunistic aloha,” Selected Areas in
Communications, IEEE Journal on, vol. 27, no. 7, pp.
1105–1119, 2009.
[30] D. Stoyan, W. Kendall, and J. Mecke, Stochastic Geometry
and its applications. Wiley, 1987.
[31] J. Nzouonta, N. Rajgure, G. Wang, and C. Borcea, “ Vanet
routing on city roads using real-time vehicu lar traffic
information,” Vehicular Technology, IEEE Transactions on,
vol. 58, no. 7, pp. 3609–3626, 2009.
[32] R. W. Yeung, S.-Y. R. Li, N. Cai, and Z. Zhang, “Network
Coding Theory Part II: Multiple Source,” Foundations and
Trends in Communication and Information T heory, vol. 2, pp .
330–381.
[33] V. Nundloll, G. S. Blair, and P. Grace, “A Component-bas ed
Approach For ( Re )-Configur able Routing in VANETs,”

38 Sabih ur Rehman et al. : Vehicu lar Ad -Hoc Networks (VANETs) - An Overview and Challen ges
Design, pp . 1–6, 2009.
[34] M . Al-Rabayah and R. M alaney, “ A new scalable hy brid rout-
ing prot ocol for vanets,” Vehicular Technology, IEEE Trans-
actions on, vol. 61, no. 6, pp. 2625–2635, 2012.
[35] A. Shrestha, “Investigation of M ANET Routing Protocols for
Mobility and Scalability,” PG Net, 2009.
[36] G. Liu, B. S. Lee, B. C. Seet, C. H. Foh, K. J. Won g, and K. K.
Lee, “A Routing Strategy for M etropolis Vehicular
Communications,” International Conference on Information
Networking (ICOIN), pp . 134–143, 2004.
[37] H. FuBler, M. Mauve, H. Hartenstein, and D. Vollmer,
“Location-Based Routing for Vehicular Ad-Hoc Networks,”
ACM SIGMOBILE Mobile Computing and Communication
Review (MC2R), vol. 7, no. 1, pp. 47–49, 2003.
[38] B. Karp and H. T. Kung, “GPSR : Greedy Perimeter Stateles s
Routing for Wireless Networks,” in Proceedings of the
ACM/IEEE International Conference on Mobile Computing
and Networking (MobiCom), 2000.
[39] V. Naumov, R. Baumann, and T. Gross, “An evaluation of
inter-vehicle ad hoc networks based on realistic vehicular
traces,” Proceedings of the seventh ACM international sy m-
posium on Mobile ad hoc networking and computing
-Mobi-Hoc ’06, p. 108, 2006.
[40] C. Lochert, H. Hartenstein, J. Tian, H. FuBler, D. Hermann,
and M. M artin, “A Routing Strategy for Vehicular Ad Hoc
Networks in City Environments ’,” in Proceedings of IEEE
Intelligent Vehicles Symposium (IV2003), vol. 2000, no. 01,
2003, pp. 156–161.
[41] C. Lochert, M. M auve, H. F¨uß ler, and H. Hartenstein,
“G eographic Routing in City Scenarios,” ACM SIGMOBILE
Mobile Comput ing and Communications Review, vo l. 9, no.
1, p. 69, Jan. 2005.
[42] V. Naumov and T. R. Gross, “Connectivity-Aware Routing
(CAR) in Vehicular Ad-hoc Networks,” IEEE INFOCOM
2007 -26t h IEEE International Conference on Comput er
Communications, pp. 1919–1927, 2007.
[43] V. Cabrera, F. J. Ros, and P. M. Ruiz, “Simulation-Bas ed
Study of Common Issues in VANET Routing Protocols,”
VTC Spring 2009 -IEEE 69th Vehicular Technology Con-
ference, pp. 1–5, Apr. 2009.
[44] C. Lin and M . Gerla, “Adapt ive clustering for mobile wireless
networks,” IEEE Journal on Selected Areas in Communica-
tions, vol. 15, no. 7, pp. 1265–1275, 1997.
[45] R. A. Santos, R. M. Edwards, and L. N. Seed, “A
Location-Based Routing Algorithm for Vehicle to Vehicle
Communication,” vol. 00, no. C, pp. 221–226, 2004.
[46] Y. Luo, W. Zhang, and Y. Hu, “A new cluster based routing
protocol for vanet,” in Networks Security Wireless Commu-
nications and Trusted Computing (NSWCTC), 2010 Second
International Conference on, vol. 1, 2010, pp. 176–180.
[47] Z. Rawashdeh and S. M ahmud, “A novel algorithm to form
st able clust ers in vehicular ad hoc networks on highways,”
EURASIP Journal on Wireless Communications and Net-
working, vol. 2012, no. 1, pp. 1–13, 2012.
[48] M. Durresi, A. Durresi, and L. Barolli, “Emergency Broad-
cast Protocol for Inter-Vehicle Communications,” 11th Inter-
national Conference on Parallel and Distributed Systems
(ICPADS’05), vol. 2, pp. 402–406, 2005.
[49] L. Zhou, G. Cui, H. Liu, D. Luo, and Z. Wu, “NPPB: A
broadcast scheme in dense VANETs,” Information Technol-
ogy Journal, vol. 9, no. 2, pp. 247–256, 2010.
[50] J. Zhao and G. Cao, “VADD: Vehicle-assist ed dat a delivery
in vehicular ad hoc networks,” in INFOCOM 2006. 25th
IEEE International Conference on Computer Communica-
tions. Proceedings, 2006, pp. 1–12.
[51] NOW -Network on Wheels , http://www.network-on-wheels.
de/, 2008,[Online ; ac cessed 20-M ay -2013].
[52] SEVECOM, “ SeVeCom (Secure Vehicular Communication
),” http://www.sevecom.org/, 2008,[Online; acc essed 20-Ma
y2013].
[53] M. T. Goodrich and R. Tamassia, ”Introduction to Computer
Security”. Pearson Education Inc.Boston, 2011.
[54] G. Calandriello, P. Pap adimitratos, J.-P. Hubaux, and A. Lioy,
“Efficient and robust pseudony mous authentication in
VANET,” Proceedings of the fourth ACM international
workshop on Vehicular ad hoc networks -VANET ’07, p. 19,
2007.
[55] D. Eckhoff, C. Sommer, T. Gansen, R. German, and F.
Dressler, “Strong and affordable location privacy in VANETs:
Identity diffusion using time-slots and swapping,” 2010
IEEE Vehicu lar Networking Conference, pp . 174–181, Dec.
2010.
[56] S. Zeadally, R. Hunt, Y. S. Chen, A. Irwin, and A. Hassan
“Vehicular ad hoc networks (VANETS): status, results, and
challen ges,” Telecommunication Systems, Dec. 2010.
[57] M. Raya and J. P. Hubaux, “The securit y of vehicular ad hoc
networks,” Proceedings of the 3rd ACM workshop on
Security of ad hoc and sensor n etworks - SASN ’05, p . 11,
2005.
[58] Nakagami, M. (1960) The m-distribution: a general formula
of intensity distribution of rapid fading, pp .3–36 in: St at istical
Methods in Radio Wave Propagation. W. C. Hoffman, Ed.,
New York: Pergamon.
[59] William Trant er, K. Shanmugan, Theodore Rapp aport, and
Kurt Kosbar. 2003. Principles of Communication Systems
Simulation with Wireles s App licat ions (First ed.). Prentice
Hall Press, Upper Saddle River, NJ, USA.
[60] D. M . Pozar, M icrowave Engineerin g, 3rd Edition. New York:
Wiley, 2010.