Cognitive Radio Enabled VANET for Multi-agent
Based Intelligent Traffic Management System
S. M. Nadim Uddin1, Nafees Mansoor2, Sazzad Hossain3
1
Department of Electronics and Telecommunication Engineering, University of Liberal Arts Bangladesh
2,3
Department of Computer Science and Engineering, University of Liberal Arts Bangladesh
E-mail: 1sm.uddin.ete@ulab.edu.bd, 2nafees.mansoor@ulab.edu.bd, 3sazzad.hossain@ulab.edu.bd
Abstract—With the mounting interest on cognitive radio (CR)
technology in wireless communication systems, it is anticipated
that CR-enabled vehicular networks will play a vigorous role in
the enrichment of communication efficiency in vehicular
network. This paper presents a Cognitive Radio enabled VANET
for multi-agent based intelligent traffic management system. A
skeleton for intelligent learning and decision mechanism for
Central Traffic Management is also proposed and discussed in
the model. The proposed model has two distinct information
exchange system layouts. One is dynamic (vehicle to vehicle) and
another is semi-dynamic (vehicle to Road-Side-Unit). For the
vehicle-2-vehicle communication, the proposed model assumes
that vehicles can communicate with each other using available
wireless resources with the help cognitive radio mechanism. This
paper also introduces a cluster formation scheme for better
accuracy in data transmission among vehicles. The dynamic
module of the proposed model is later simulated and validated
for some important performance communication metrics.
Keywords—Cognitive Radio; VANET; Traffic Management
System; CR-VANET; MANET
I.
INTRODUCTION
With the advancement of microelectronics and
communication modules, a rapid surge of interest is observed
in the research community for Mobile Ad-hoc Network
(MANET) and Vehicular Ad-hoc Network (VANET).
MANET, a special type of ad-hoc network, is self-organized
and is fully operable without the assistance from any fixed
infrastructural support or central administration. The mobile
nodes are inter-connected by the wireless links and
communication is held directly between nodes or through
intermediate nodes [1]. Each node in MANET continuously
maintains the updated information that is required to properly
route the traffic [2]. The Internet Engineering Task Force
(IETF) has developed two standard track routing protocols for
MANET, namely proactive and reactive MANET protocols
[3].
On the other hand, VANET is a distinctive class of
MANET where moving vehicles act either as nodes for direct
communication or as routers to provide intermediate
connectivity. These vehicular nodes can communicate with
other vehicles to establish Vehicle-to-Vehicle communication
system. These nodes can also communicate with the access
point (AP) to establish an Infrastructure-to-Vehicle
communication system. A VANET network consists of four
major components, namely Vehicles (nodes or mobile hosts),
On-Board Unit, Road-Side Unit and Central management
system [5].
Thus, VANET follows and applies the same principle of
MANET in a highly dynamic environment of surface
transportation. However, due to the high mobility of the
nodes, VANET needs to consider dynamic information
exchange and unreliable channel conditions. These
considerations are absent in MANET. As a result, deployment
of MANET protocols in VANET show poor performance [6].
Since the vehicles move along the road, movements of the
vehicles are predictable in VANET, where network density
changes over time and location [4].
According to the IEEE 802.11 independent basic service
set (IBSS), no access point is required to communicate in
distributed peer-to-peer manner and IBSS operation can occur
if two nodes are within the radio range of each other. [5].
However, the radio spectrum scarcity has become more
vigorous concern with the amplifying demand in wireless
applications. To combat the growing demand of radio
spectrum, proper utilization of the radio spectrum is essential.
Cognitive radio exercises dynamic spectrum allocation
technique to utilize radio spectrum efficiently and reduces the
bottleneck on frequency bands. With the recent advances in
cognitive radio systems [23-24], cognitive radio enabled
vehicular users in VANETs would be able to sense and hop
from one frequency to another (or one network to another) in
the entire spectrum range based on their needs and operating
environment with the help of cognitive radios. While several
studies exist in literature on applying CR to wireless mesh
networks, ad hoc networks, and cellular net- works, the
research on applying CR to VANETs is still in its early stage.
The research solutions proposed for general-purpose CR
networks cannot be directly applied to CR-VANETs as the
unique features of vehicular environment, such as the role of
mobility, and the cooperation opportunities need to be taken
into account while designing the spectrum management
functions for CR-VANETs [25-26].
In this paper, Cognitive Radio Enabled VANET for multiagent based intelligent traffic management system is proposed.
The proposed system ensures data exchange between highspeed vehicles and between the vehicles and the roadside
infrastructure in the licensed ITS band (5.85-5.925 GHz). In
this model, vehicles can communicate with each other using
available wireless resources with the help cognitive radio
mechanism. Without the loss of generality, the proposed
model is simulated and validated for some important
performance communication metrics. In this paper, a cluster
formation mechanism is considered for better accuracy in data
transmission among vehicles. A skeleton for intelligent
learning and decision mechanism for a Central Traffic
Management is also proposed and discussed in the scope of
the model.
The paper is organized as follows. In section II, a brief
analysis on different lately developed protocols for VANET
Communications is discussed. The proposed network model
and architecture for “Cognitive Radio Enabled VANET for
Multi-agent Based Intelligent Traffic System” are described in
section III. In section IV, the simulation results of the
proposed model are presented. Conclusion and future works
have been discussed in section V.
II.
RELATED WORKS
Different protocols have been proposed for better
performance
in
wireless
networks
for
vehicular
communications. OLSR sends two types of messages namely
hello message and Topological control message [7]. This
protocol gives better performance among the proactive routing
protocols [8]. DSDV uses the shortest path to find the route to
the destination and guarantees loop free nodes reduces count
to infinity problem and also reduces control message
overhead. This protocol is suitable only for smaller number of
nodes [9].
Ad-hoc on-demand distance vector (AODV) protocol
discovers routes only on demand i.e. it establishes a route only
when any node needs to send a message to the destination. It
offers low network overhead by avoiding the flooding of
messages periodically in the network. It requires less memory
size and the routing tables only contain the recent active
nodes. AODV is flexible to highly dynamic and large-scale
network [10]. Ad-hoc On-demand Multipath Distance Vector
Routing (AOMDV) protocol maintains multiple loop free path
with minimum overhead. It is suitable for high mobility nodes
[11]. Dynamic Source Routing (DSR) protocol mainly
consists of two mechanisms, namely route discovery and route
maintenance and uses a unique id request in the route request
packet [13]. Temporally Ordered Routing Algorithm (TORA)
protocol uses multi hop routes. This protocol is based on the
link reversal routing algorithm which uses directed acyclic
graph to identify the flow of packets . TORA’s performance is
better than DSR in highly dynamic ad-hoc environment [13].
Zone Routing Protocol (ZRP) is the first hybrid routing
protocol which divides the network into overlapping zones. It
uses the proactive routing scheme inside the zone and reactive
routing scheme outside the zone [14]. Core Extraction
Distributed Ad-hoc Routing (CEDAR) is a protocol with
integrated QoS support [15]. Distributed Dynamic routing
algorithm Protocol (DDR) is a tree based routing protocol that
does not require the root node support for data transfer.
Greedy Perimeter Stateless Routing (GPSR) selects node
closer to the destination using beacon [17]. Greedy Perimeter
Coordinator Routing (GPCR) is a position-based overlay
routing protocol that uses greedy algorithms to forward packet
based on a pre-selected path .It has been designed to meet the
challenges of city scenarios. No Global Information System
required for GPCR [18]. Connectivity- Aware Routing (CAR)
is well suited for city and highway scenarios. It uses AODV
for path discovery and PGB for data dissemination. It also
uses guard concept for path maintenance. It ensures the
shortest connected path and no digital map is required for
CAR. It has higher packet delivery ration than GPSR [18].
GSR (Geographic Source Routing) is designed for city
environment, uses greedy forwarding approach along preselected path using Djkshtra’s shortest path algorithm. It
combines the features of topological information and position
based routing [19]. Recent cognitive radio based wireless
mesh network related work either use the channel selection
within a network [10] or transmit power or rate adaptation
within a given network [11]. These types of solutions are not
directly applicable in heterogeneous wireless environment
since different networks have different characteristics. All of
the aforementioned research works are related to protocol
designing for VANET & protocol proposals and spectrum
sensing for CR.
However, there is no cluster-based ad-hoc routing
protocol-integrated intelligent surface traffic management
system has been proposed.
III.
PROPOSED MODEL
A. Network Model
In the proposed model, it is assumed that there are two
distinct scenarios in terms of nodes’ density namely highly
dense network and light dense network. A mobile vehicle is
noted as node where mobile vehicles can be cars, buses etc. in
the network. A sender node is defined as a particular node
from where the data packets are coming and destination node
is defined as the sink or the desired recipient of the data
packets. A grid is defined as a geographical area with at least a
Roadside Unit (RSU) and a predefined number of nodes with
On-Board Units (OBU) installed into them. Here node density
is the number of nodes in a grid and network is considered as
the accumulation of all grids. A RSU is a hardwire mounted
on top of a pole or tower in a grid, which is capable of
maintaining simultaneous wireless duplex connections with
nodes and Central Traffic Management Unit (CTMU). RSUs
build a neural network with each other and the CMTU.
In a grid, there are two types of networks, which is labeled
as Dynamic networks and Semi-dynamic networks. Dynamic
network is created among nodes where nodes can move along
a predefined manner and semi-dynamic network is created
between a node and a RSU.
In a dynamic network, the nodes can communicate with
each other using the principle of cognitive radio with the help
of OBUs, which follow the IEEE 802.11p protocols. OBUs
are connected with each other via wires or wireless networks.
CMTU is the central management system, which periodically
updates traffic conditions and sends relevant data to RSUs for
proper route maintenance and nodal information to be
displayed in Dynamic Info Board (DIB). A DIB is an info
board from where nodal position and approximate speed and
arrival time can be seen. It is mounted on the same pole of a
RSU.
Data packet, in the model, follows a basic structure
consisting of sender node id, destination id, hop count, timer,
sequence number etc. The routing path for source to
destination is determined by the greedy forwarding algorithm
which means node closest to the destination node among all
neighboring nodes in the transmission range of the sender
node is selected as the next relaying node or the next hop. It is
assumed that MAC protocol is TDMA, in which time is
slotted and synchronized and to ensure proper sharing of the
wireless resources, an appropriate scheduling algorithm is
selected. For simplicity, rerouting mechanism (in case of link
failure), packet collision probability (two nodes trying to send
data packets to one node at the same time), nodes' mobility
(speed) etc. are not considered in the scope of this paper.
approximate arrival time in DIB. An overview of advanced
traffic system is shown in Fig 1.
Suppose a node “A”, which is a local transport, enters a
grid namely “XYZ”. In XYZ, there are several nodes “B”,
“C”, “D”, “E” etc. “A” is then connected to a network label as
“GHJ123” in which all other nodes are connected as well. In
that grid, there is a RSU, labeled as “RSU1”, at an
intersection. “A” needs to send data packets to RSU1, but it is
not in the range of “A”. So, the data packets from A can be
sent to RSU1 via B, C and D. RSU1 receives the data packets
and send it to the CTMU. CMTU processes data and sends
the updated data to the RSU1, RSU2, RSU3 etc. A basic
network with links among nodes and RSU is shown in Fig 2.
B. Architecture:
The proposed model is an intelligent traffic management
system mainly focused on city scenarios though it can be
implemented in highways as well as rural areas.
In an arbitrary geographical location, it is assumed that
there are some grids with some fixed RSUs. When a node
enters a grid, it is automatically connected to the wireless
network of that grid through its OBU. OBU then sends data
packets to its nearby nodes or RSU. In a wireless ad-hoc
network, a node can only send data packets if there is another
node or the desired node within the transmission range of the
sender node.
If the RSU is not in the range of the sender node, the data
packets can be sent from the sender node to the destination
node or RSU via intermediate nodes, which are labeled as
relaying nodes.
Fig 2: Node view
According the mechanism discussed in the Network Model
subsection, clusters form among the nodes to achieve better
accuracy in data packet exchange. In a cluster, all links among
the nodes have link-weight, which depend on the possession
of available channels in cognitive networks. The node with
greater link-weight is selected as the "Cluster head" and other
nodes within the transmission range of the CM become
members of that particular cluster. Cluster-head becomes
responsible of transmitting data packets to the destination or to
the next cluster through "Edge Members". Edge members are
responsible for maintaining the links among clusters. To avoid
data packets loss, a "Secondary Cluster head" is selected in
case of unavailability of Cluster head. An example of cluster
formation among nodes is illustrated in Fig 3.
Fig 1: The proposed traffic management system
A RSU picks up the data packets sent to it and extract the
necessary information for traffic conditions in that grid and
sends the data to CTMU for further processing. CMTU then
processes the received data and sends the updated data to the
RSUs. RSUs then display information of selected vehicles
(local transports, ambulance etc) such as positions and
Fig 3: Proposed clustering scheme
On-Board Unit is a hardware implemented on the vehicles,
which creates and maintains links with other nodes or RSU. It
operates according to IEEE 802.11 protocols and it utilizes the
available wireless resources such as bandwidth, channels etc.
with the help of cognitive radio network mechanism. Though
the OBU will continuously consume energy as it is active as
long as the vehicle is active, it is assumed that the node has
sufficient power supply for it. The main task of the OBU is to
sense spectrums and maximize the possibility to be online by
using or sharing available bandwidth so that it can broadcast
its data packets to the destination node.
A DIB is a board, which displays some basic real time
information about the partial grid traffic and local transports.
It is connected to the RSU on which it is mounted. Whenever
the RSU is updated from the CMTU, updated information
about some specific vehicles shows up on the DIB. It shows
approximate location, time of arrival, accidents etc.
CMTU is the main server or the accumulation of several
main servers, which controls the whole traffic system. It is
connected with the RSUs via wire or wireless connections and
builds a neural network. CMTU has an intelligent learning and
decision-making algorithm, which extracts necessary data
from the data, received from the RSUs and updates its
database for future decision making. CMTU then updates the
RSUs and the specific updated data for specific route is shown
in the DIB of specific RSU. CMTU allocates time intervals in
intersections based comparison in previous data and the
updated data in its database.
C. Cluster formation
With the mounting interest on cognitive radio (CR)
technology in wireless communication systems, it is
anticipated that CR-enabled vehicular networks may improve
the vehicular communication efficiency. Thus, considering the
possibilities of cognitive radio technology, the dynamic
module of the proposed network model is designed to be
functional on cognitive radio environment. As discussed
earlier in the network model section, the dynamic vehicular
network in each grid is divided into some sub-groups or
clusters. Clustering concept is introduced in the dynamic
module as cluster-based ad-hoc network aims to achieve better
accuracy in data packet exchange.
The proposed clustering mechanism is inspired from
author’s previous clustering scheme for cognitive radio ad-hoc
network [20-22]. In the existing clustering mechanism,
cluster-head selection is based upon a weight, where to
calculate the weight number of common channels and number
of neighboring nodes is taken into consideration. However, in
this paper another parameter has been taken into account
called node’s speed along with the previous two to calculate
weight for each node.
Once the weight calculation for each node is completed,
the node with higher weight is selected as the "Cluster head"
and other nodes within the transmission range of the Clusterhead (CH) become cluster members of that particular cluster.
Cluster-head becomes responsible of transmitting data packets
to the destination or to the next cluster through "Edge
Members". Edge members are responsible for maintaining the
links among clusters. To avoid data packets loss, a "Secondary
Cluster head" is selected in case of unavailability of Cluster
head. An example of cluster formation among nodes is
illustrated in Fig 3.
In the proposed clustering scheme, CH defines and
upholds operating channels for the cluster. To find the
existence of any other clusters in the neighborhood, CMs
check their neighbor list for other cluster heads. CM becomes
the Edge Member (EM) and connects two clusters once it
finds other CH in the neighbor list. In the proposed clustering
scheme, cluster consists of one CH, one SCH and CMs. All
cluster members are 1-hop apart from the CH. EM connects
two neighboring clusters, where there can be maximum two
intermediate EMs between two CHs. Using local common
channels, intra-cluster communications are performed.
IV.
SIMULATION RESULTS
A. Simulation Environment
To evaluate the clustering performance of the dynamic
module (vehicle-2-vehicle) of the proposed model, simulation
is conducted. Though several network simulators are
available, whose output depicts as close as possible to real
time implementation, to simulate and analyze performances of
the proposed model, discrete-event simulator NS2 has been
used and the performance analysis are conducted using PERL
scripts.
Four parameters, namely throughput, energy consumption,
delay and overhead are considered as the performance metrics
to evaluate the performance of the network. Moreover, two
distinct dynamic scenarios in terms of nodes’ density are
considered for the comparative study in the simulation
environment. Thus, in one scenario the radio transmission
range of a node is considered to be 100 meters and in the other
scenario, the transmission range is set to 500 meters. For both
scenarios, number of nodes is considered to be 100 where
distributed sources and sinks are altered randomly.
A simulation area of 4000 m² is considered for the
simulation purpose. The Two-Ray Ground model is used as
the propagation model and Drop-tail method is used for the
queuing purpose. MAC/802.11p is considered as the MAC
type and maximum packet queuing delay is considered as
50ms. In the simulation, data traffic is generated with Constant
Bit Rate (CBR) with packet size is set to 512 bytes. Varied
packet rate ranging from 100 packets/sec to 800 packets/sec is
considered to evaluate the performance of the network for
different traffic load. Initial energy for all the nodes is
considered to be 10 Joules. Maximum speed for the nodes is
considered to be 10 m/s. The simulations are run for 150s each
and the results are calculated as mean of several observations
in light dense dynamic environment and heavily dense
dynamic environment.
B. Performance Evaluation
This section of the paper discusses the simulation results of
the proposed model in terms of throughput, energy
consumption, delay and overhead.
1.
packets. In Fig. 5, the horizontal axis indicates the traffic load
and the vertical axis indicates the consumed energy.
Performance based on Throughput
In this paper, throughput is defined as the number of
successfully received data packets at the destination node in a
unit time and it is represented in Kbps. Maximum throughput
is preferable while designing the routing protocol for the
proposed dynamic network. In Fig. 4, the horizontal axis
indicates the traffic load and the vertical axis indicates the
throughput.
Fig. 5: Performance Evaluation in Terms of Consumed Energy
From the figure, it is seen that the energy consumption
increases with increasing data flow rate in both scenarios. The
reason is that when traffic load is increased, more data packets
need to be transmitted over the network from the source node
to the destination node, where the intermediate nodes are
required to be at the active state for longer period of time. As
consumed energy of a node is considered to be dependent on
the functional period, therefore with increasing traffic load,
energy consumption is also increased in both networks.
Fig 4: Performance Evaluation in Terms of Throughput
Fig. 4 shows that with an increasing traffic load,
throughput for both scenarios increases. However, throughput
is higher in the network with the radio transmission range of
500 meters, than the radio transmission range of 100 meters
for all different data flow rates. This is because; a network
with long ranged transmission finds lesser number of hops to
transmit packets to the destination. Thus, with decreasing
number of hops, number of links throughout the network,
probability of link failure and rate of packet retransmission
reduce significantly. Therefore, with higher transmission
range, throughput of the network increases for all different
traffic loads.
2.
Moreover, it can also be seen from Fig. 5 that the energy
consumption is lesser in a network with radio transmission
range of 500 meters compared to the network with radio
transmission range of 100 meters for all different traffic loads.
This is because, when radio transmission range in a network is
longer, number of hops from the source to the destination
tends to be lesser than that of a network of shorter
transmission ranged radios. Thus, a network with long ranged
transmission results lesser number of intermediate nodes
between a source and the destination than the network with
short ranged transmission. Therefore, energy consumption
remains lesser in a network with longer transmission ranged
radios as lesser number of intermediate nodes is engaged to
forward the data packet from the source node to the
destination.
Performance based on Energy consumption
3.
In this paper, energy consumption is defined as the
cumulative sum of consumed energy by all the nodes of the
network during the entire transmission period, where
consumed energy of a node is calculated by subtracting
remaining energy from initial energy. The unit for energy
consumption has been considered as Joule. Minimum energy
consumption is desirable while designing the routing protocol
for the proposed dynamic network. In the simulation
environment, consumed energy of a node mainly depends on
the functional period of a node to transmit the data packets to
its next hop. Moreover, network energy consumption also
depends on the number of relying nodes while transmitting
Performance based on Packet Transmission Delay
In this paper, the packet transmission delay is defined as
the average time required for transferring data packets from
the source node to the destination node. Minimum packet
transmission delay is preferable while designing the routing
protocol for the proposed dynamic network. In Fig 6, the
horizontal axis indicates the traffic load and the vertical axis
indicates the delay, where it is seen that the delay varies with
varying data flow rate in both scenarios.
From the figure it is observed that the packet transmission
delay increases with increasing data flow rate in both
scenarios. The reason is that when traffic load is increased,
more data packets need to be transmitted from the source to
the destination node. As a result, the intermediate nodes are
required to process more packets which eventually increases
individual data processing sessions among the nodes. Thus,
when higher number of packets propagates, source node and
the intermediate nodes need longer time to forward the packets
to the next hop, which increases the cumulative packet
transmission delay.
transmitted control packets over total received data packet at
the destination node. In Fig. 7, the network overhead ratio is
presented in terms of percentage, where network overhead
ratio is desirable to be minimum while designing the routing
protocol for the proposed dynamic network. In the figure, the
horizontal axis indicates the traffic load and the vertical axis
indicates the network overhead ratio.
Fig. 6: Performance Evaluation in Terms of Packet
Transmission Delay
From the figure, it is seen that the overhead ratio increases
with increasing data flow rate in both scenarios. The reason is
that when traffic load is increased, more data packets need to
be transmitted over the network from the source node to the
destination node. In a wireless ad-hoc network, each node has
a defined data packet queuing delay. Increase in data rate
means nodes need to process more data packets individually
which eventually results in reduction of a node's efficiency for
packet forwarding. As a result, with the increase of traffic
load, retransmission of data packets from source node to
destination node increases and therefore, acknowledgement
control messages for retransmission to source node and from
destination node increase, network overhead increases. Thus,
with the increase in traffic load, the network overhead ratio
increases.
Moreover, it can be also observed from Fig. 6 that the
packet transmission delay is lesser in a network with radio
transmission range of 500 meters than in the network with
radio transmission range of 100 meters for all different traffic
loads. This is because, when radio transmission range in a
network is longer, number of hops from the source to the
destination node is reduced than that of a network with shorter
transmission ranged radios. Thus, lesser number of
intermediate nodes is engaged to forward the data in the
network, which results lesser data processing sessions.
Therefore, a network with long transmission ranged radios
results lesser packet transmission delay than that of a network
with short transmission ranged radios.
Moreover, it is also observed from Fig. 7 that the overhead
ratio is lesser in a network with radio transmission range of
500 meters compared to the network with radio transmission
range of 100 meters for all different traffic loads. That is
because, nodes in a network with shorter radio transmission
range will need more number of relaying nodes for
transmitting data packets to the destination node from the
source node than that of with longer radio transmission range.
As the number of the nodes is comparatively more, more
nodes will have to process increased traffic load and as a
result, retransmission of data packets due to exceeding node's
capacity to process data packets will increase which
eventually increases number of transmitted control packet.
Thus, a network with radio transmission range of 100 meter
will have more network overhead ratio than that of a network
with radio transmission range of 500 meters.
V.
Fig. 7: Performance Evaluation in Terms of Overhead Ratio
Performance based on Overhead Ratio
The network overhead is considered as the sum of
transmitted control packets during the transmission sessions.
Thus, network overhead ratio is defined as the total
CONCLUSION AND FUTURE WORKS
This paper presents a Cognitive Radio enabled VANET for
multi-agent based intelligent traffic management system. The
proposed model has two distinct information exchange system
layouts, namely dynamic module (vehicle to vehicle
communication using cognitive radio) and semi-dynamic
module (vehicle to Road-Side-Unit). For the dynamic module,
a cluster formation scheme is introduced and later simulation
is conducted to validate the performance. Our next research
steps are to develop the prototype of the model and to develop
an intelligent algorithm that can give optimal decision to
manage the traffic.
REFERENCES
4.
[1]
Sesay, S., Yang, Z., & He, J. (2004). A survey on mobile
ad hoc wireless network. Information Technology
Journal, 3(2), 168-175.
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
Timcenko, V., Stojanovic, M., & Rakas, S. B. (2009,
August). MANET routing protocols vs. mobility models:
performance analysis and comparison. In Proceedings of
the 9th WSEAS international conference on Applied
informatics and communications (pp. 271-276).
Retana, A. Mobile Ad-hoc Networks, charter-ietf-manet03 (1997)
Anwer, M. S., & Guy, C. (2014). A survey of VANET
technologies. J. Emerg. Trends Comput. Inf. Sci, 5, 661671
Nadeem, A., & Howarth, M. P. (2013). A survey of
MANET intrusion detection & prevention approaches for
network layer attacks.Communications Surveys &
Tutorials, IEEE, 15(4), 2027-2045.
Li, F., & Wang, Y. (2007). Routing in vehicular ad hoc
networks: A survey.Vehicular Technology Magazine,
IEEE, 2(2), 12-22
Bilal, S.M., C.J. Bernardos, and C. Guerrero, Position
Based Routing in Vehicular Networks: A Survey. Journal
of Network and Computer Applications, 2012.
Philippe Jaquet, Paul Muhlethaler and Amir
Qayyum,―Optimized Link State Routing Protocol‖,
IETF
Jamal
toutouh,Jose
Garcia-Nieto
and
Enrique
Alba,Intelligent OLSR routing Protocol Optimization for
VANETs,IEEE
transactions
on
vehicular
technology,vol.61,No.4,May 2012
Bai, Y., S.B. Ji, and Y. Xue, A Comparative Study of
AODV and DSDV Routing Protocol of the Vehicle Ad
Hoc Network Based on NS2.
C. E. Perkins and E. M. Royer, ―Ad Hoc On-demand
Distance Vector Routing‖, In Proceedings of the 2nd
IEEE Workshop on Mobile Computing Systems and
Applications, New Orleans, LA, 1999, 90-100.
D. B. Johnson, D. A. Maltz, Yih-Chun Hu and Jorje
tagJetcheva, ―The Dynamic Source Routing for Mobile
Ad Hoc Wireless Networks‖, IETF Internet draft,
Nov.2001.
Pirzada, A.A., C. McDonald and A. Datta,
2006.Performance comparison of trust-based reactive
routing protocols. Mobile Computing, IEEE Transactions
on, 5(6): 695-710.
Kashif Naseer Qureshi and Abdul Hanan Abdullah ,
Study of Efficient Topology Based Routing Protocols for
Vehicular Ad-Hoc Network Technology, World Applied
Sciences Journal 23 (5): 656-663, 2013
Karp, B. and Kung, H. T (2000), ―GPSR: greedy
perimeter stateless routing for wireless networks. In
Mobile Computing and Networking, pages 243-254,
2000.
Naumov, V., Baumann, R., Gross, T. (2006), ―An
evaluation of Inter-Vehicle Ad Hoc Networks Based on
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
Realistic Vehicular Traces, Proc. ACM MobiHoc’06
Conf., May, 2006.
Lochert, C., Mauve, M., F¨ussler, H., and Hartenstein, H.,
―Geographic routing in city scenarios,‖ SIGMOBILE
Mob. Comput. Commun. Rev., vol. 9, no. 1, pp. 69–72,
2005.
Naumov, V., Gross, T.R. (2007), "Connectivity-Aware
Routing (CAR) in Vehicular Ad-hoc Networks,"
INFOCOM 2007. 26th IEEE International Conference on
Computer Communications. IEEE , vol., no., pp.19191927, 6-12 May, 2007.
M. Jerbi, S. M. Senouci, R. Meraihi, Y. Ghamri-Doudane,
―An improved vehicular ad hoc routing protocol for city
environments, In Communications, 2007, pp. 3972-3979.
Mansoor, N., Islam, A. K. M., Zareei, M., Baharun, S., &
Komaki, S. (2013, December). Spectrum aware clusterbased architecture for cognitive radio ad-hoc networks. In
Advances in Electrical Engineering (ICAEE), 2013
International Conference on (pp. 181-185). IEEE.
Mansoor, N., Islam, A. M., Zareei, M., Baharun, S.,
Wakabayashi, T., & Komaki, S. (2015). Cognitive radio
ad-hoc network architectures: a survey. Wireless Personal
Communications, 81(3), 1117-1142.
Mansoor, N., Islam, A. K. M. M., Zareei, M., Baharun, S.,
& Komaki, S. (2014). Cluster modelling for cognitive
radio Ad-hoc networks using graph theory. In 2014
International Conference on Applied Mathematics,
Modelling and Simulation (ICAMMS) (pp. 1-8).
G. Marfia, M. Roccetti, A. Amoroso, M. Gerla, G. Pau,
and J. H. Lim, “Cognitive Cars: Constructing a Cognitive
Playground for VANET Research Testbeds,” in 4th
International Conference on Cognitive Radio and
Advanced Spectrum Management (CogART 2011,
Barcelona, Spain, Oct. 2011
D. B. Rawat, B. B. Bista, and G. Yan, “CoRVANETs:
Game Theoretic Approach for Channel and Rate
Selection in Cognitive Radio VANETs,” in Proceedings
of the Seventh International Conference on Broadband
and Wireless Computing, Communication and
Applications (BWCCA-2012), Nov 2012
Yan, M. Li, Z. Yang, W. Lou, and H. Zhai, “Throughput
Analysis of Cooperative Mobile Content Distribution in
Vehicular Ad Hoc Networks using Symbol Level
Network Coding,” IEEE Journal on Selected Areas in
Communications (JSAC), special Issue on Cooperative
Networking: Challenges and Applications, 2012.
P. Shankar, T. Nadeem, J. Rosca, and L. Iftode, “CARS:
Context-Aware Rate Selection for vehicular networks,” in
Proceedings of the IEEE International Conference on
Network Protocols, 2008. ICNP 2008., November 2008,
pp 1-12