www.ijecs.in
International Journal Of Engineering And Computer Science ISSN: 2319-7242
Volume 5 Issue 5 May 2016, Page No. 16383-16399
Routing Taxonomies for Network Aware Territory of Vehicles: A Review
T. Sivakumar 1, Ali Tauseef Reza2, T. Anil Kumar 3
1
Department of Computer Science, School of Engineering & Technology,
Pondicherry University, Puducherry, India
tsivakumar@yahoo.com
2
Department of Computer Science, School of Engineering & Technology,
Pondicherry University, Puducherry, India
alitauseefreza@gmail.com
3
Department of Computer Science, School of Engineering & Technology,
Pondicherry University, Puducherry, India
anilkumar.ak1991@gmail.com
Abstract: In this task we aim to provide a through global taxonomy of VANET routing protocols. This task also aims to provide a
simulation test bed enabling performance assessment of the protocols. This work also complements the previous approaches of
classification. Characteristically we acknowledged following taxonomical routing protocols classification, based on their transmission
approach, based on their prerequisite knowledge needed to realize routing, based on their delay sensitivity and toleration, based on their
accommodating network i.e., heterogeneous and homogeneous vehicular network environment and based on their inspiration i.e., bioinspired algorithms. Evaluation of a routing protocol in VANET is a necessary, indispensable and struggling task, so we bring assessment
methods, i.e., simulation and real world research into the picture. Once the protocol passes all the simulation tests with expected results then
it can be tested in the real time vehicular environments. All of this work provides a base for VANET research community to excogitate a new
routing techniques.
Keywords: VANET, Routing Protocol, IVC, ITS, WAVE, DSRC.
1. Introduction
VANETs was considered as an offshoot of MANETs but has
now become a special area of research. Fundamentals of
MANETs – the unplanned and voluntary formation of wireless
network for data exchange – are applied to VANETs.
Particulars and technicalities of MANETs and VANETs
concede as well as contradicts. There are various similarities
and dissimilarities between MANETs and VANETs.
Eventually VANET has now ripen into a crucial part and parcel
of Intelligent Transportation System (ITS). ITS exploits new
technologies to minimize the road fatalities and maximize
road‟s efficiency. [1] - [3] provide a comprehensive study of
problems and expenses aroused because of increasing no of
vehicles.
Dealing with routing take up utmost priority in giving values
to VANETs applications in Inter Vehicular Communications
(IVC), ITC, etc. In VANET scenario routing deals with the
techniques, practices and procedures of choosing optimal
journey among the paths available between packet‟s origin and
destination vehicle. Less routing overhead, delay and high
message delivery proportion are principal metrics in judging
routing protocols efficiency. In VANETs high-speed of
vehicles causes continuously changing network topology,
process of route finding to be delayed and data packet to be
lost. Because of inherent complication of VANETs there has
been great passion among researchers to art a sound and
effective routing protocols that for IVC, ITC, etc. applications.
Fig. 1. VANET architecture depicted by C2C communication
consortium [4].
VANETs can use any wireless networking technologies as
their basis of communication in V2V or V2I mode as shown in
Fig. 1 [4]. The most projected technology is Dedicated ShortRange Communication (DSRC) acknowledged as IEEE
802.11p Wireless Access in Vehicular Environment (WAVE).
Fig. 2, shows the relationship between IEEE 1609 (WAVE)
and IEEE 802.11p. Auxiliary technologies being used are
WiMAX IEEE 802.16, Bluetooth IEEE 802.15.1, MBWA
IEEE 802.20, ZigBee IEEE 802.15.4, Infrared and wireless.
The main units of WAVE are On Board Unit (OBU), Road
Side Unit (RSU) and Application Unit (AU). DSRC and
WAVE are standards proposed for VANET routing. Upper
layers of WAVE are being supported by IEEE 1609 family of
standards.
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16383
DOI: 10.18535/ijecs/v5i5.12
In the last few years many VANET routing protocols review
works, has been done but a thorough review work was felt with
simulation and evaluation test bed. A routing protocol suitable
in one application scenario may not be suitable in other. A
routing scheme could have more than one objective.
Though the VANET routing protocols classification
received their due, but the ideas associated with previous works
are so recondite that a thorough classification is required – the
main motivation behind this work.
may use intermediate wireless nodes either by using
opportunistic technique, buffering/carry-and-forward strategy
or greedy forwarding technique. Greedy forwarding strategy
sends packet to the far-most neighbor in the planned direction
while buffering technique may hold a packet until a forwarding
opportunity is available. Classically routing implies unicast
routing which can be grouped according to Fig. 3, to be taken
up in next sub-sections.
Fig. 2. IEEE 1609 (WAVE) architecture and relationship to
the IEEE 802.16p MAC and physical layers
Fig. 3. Unicast Routing Protocol Taxonomy
The organization of this work goes as follows. Section II
discusses routing taxonomies. Various distinct aspects of
classification is considered, i.e., transmission approach,
knowledge needed, delay awareness, motivating inspiration and
accommodating network. Performance assessment and
simulation test bed are discussed in section III. Finally, section
IV presents the conclusions and future orientation for the
design, research and reasoning of routing protocols.
2. Routing Classification
Driver‟s helping hand, crash alert and collision prevention
are few objectives of ITS safety applications provided to
vehicles which form a VANET and it requires routing the
packets between source and intended vehicles. Safety
applications offered by fixed infrastructure vehicular networks
such as Road-Weather Management, Crash Prevention and
Alert, Freeway Management and Safety depends on safety
information to be disseminated at right time. A lot timeprivileged
routing protocols like Dynamic Source Routing
(DSR), Destination-Sequenced Distance-Vector (DSDV),
Optimized Link State Routing Protocol (OLSR) and Ad-hoc
On Demand Distance Vector (AODV) are changed from the
MANET study. Geographic i.e., position based protocols like
GPSR and then GPCR [5] were thought by researchers for
frequent topology changing networks. Further, this paper
describes a detailed classification of the various protocols
based on some aspects mentioned in introduction of this paper
Section I. We also include the timeline of the algorithms. Most
of the protocols classified are applicable in a
particular/common scenario and to a limited scale.
A. Based on Transmission Approach
1) Unicast Routing Protocols:
These protocols refers to packet routing from a one source to
a one destination. In between source to destination protocols
Topology-based routing is assessed as the conventional style
of routing packets as in MANETs. Commonly the topologybased routing makes bookkeeping of links information. Further,
they are split up on as proactive and reactive protocols.
Vehicles sends route discovery packets only on demand in case
of reactive routing while repeatedly sends route discovery
packets at regular intervals in case of proactive routing. AODV
and DSR are renowned reactive protocols while OLSR and
DSDV are proactive.
Geographic routing is more suitable in VANET environment
as the algorithms uses the whereabouts of the source and
destination nodes. Nodes involved in forwarding are familiar of
their neighborhood. Position/path/map based routing stand in
need of GPS.
2) Multicast/Geocast Routing Protocols:
a) Geocast:
Geocast refers to the transmission of packets to a batch of
vehicles in vehicular network identified by their geographical
locations called Zone of Relevance (ZOR). A geocast-message
is only meaningful to a vehicle if it is in ZOR i.e., the vehicle
necessarily meet a set of geographical/topographical norms.
Another concept apart from ZOR is of Zone of Forwarding
(ZOF) which is a set of geographical/topographical norms a
vehicle necessarily meet to dispatch geocast messages. An
instance of ZOR in VANET is shown in Fig. 4. As a contrast to
pure-flooding based protocols geocast protocols allows
flooding of messages only in ZOF thus causing less network
congestion. Further, Geocast algorithms are split up on as
Beaconless-based and Beacon-based protocols. Fig. 5, shows
the taxonomy of geocast routing protocols.
Bachir et al. [6] came up with the Inter Vehicular Geocast
(IVG) protocol for warning every vehicle of a roadway in case
of barricade, hindrance and danger because of collision,
accident or casualty. GPS helps in determining vehicles moving
direction, its position and its velocity, these parameters are
used by IVG algorithm for defining multicast group (risk areas)
dynamically and temporally. The scenario of relay selection
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16384
DOI: 10.18535/ijecs/v5i5.12
after an accident on highway has been depicted in Fig. 6. The
vehicle which has met with accident starts to broadcast an
alarm/emergency message and the way by which a node is
nominated as relay is established on the defer-time, which is
given by the equation,
Defer_time(x)= MaxDeferTime.
where BOd , Rtx , d , MaxBOd and Sd are the backoff time
depending on the separation, the effective communication radio
radius, the separation of current and last transmitter node, the
maximum back off time and the distance sensitivity factor used
respectively.
(R - Dsx )
R
Where R is radio radius and Dsx is the separation between
the vehicle „s‟ and „x‟.
Fig. 5. Geocast routing taxonomy based on relay selection.
Fig. 4. ZOR, Multicast-group in Geocast Routing.
Following algorithm is executed by each relay node.
begin
when an emergency packet received by node(x)
if m is not important
then discard it
else node(x) defer-time has been set
when the timer ends it rebroadcast the packet
end
As IVG uses GPS for packet delivery to risk areas which
uplift the costly operations for dense-dynamic situation for
maintaining multicast tree such as neighbour estimation and
routing. The simulation outputs done with Glomosium in [6]
showed that the IVG protocol is scalable and reliable.
Maihöfer et al. [7] came up with the idea of caching
unforwardable messages which a node cannot forward because
of network segregation and torublesome neighbors. Having
small cache at the network layer decreases network load and
lag. Beaconing subsystem is used to be able to find the
information about neighbour nodes. The cache holds the table
of neighbour nodes and whenever there is any change among
the neighbour nodes then the cached messages is scanned to
find whether there is any packet which can be delivered
according to changed table configuration.
Joshi et al. [8] came up with the idea of Distributed Robust
Geocast protocol (DRG) which uses the distance-based backoff
algorithm on the , more distant is more reliable, principle of
selecting relay nodes. Its algorithm maintains restricted and
directed flooding of messages and is completely distributed and
stateless so the overhead is minimal and at the same time
simulation results shows that its reliability is equal to the fullflooding protocols. Among the contending nodes the farthest
node gets the chance as the babackoff time is conversely
proportional with the last sender distance. Formula for
calculating the backoff time based on distance is
Rtx d
BOd Rtx, d MaxBOd .Sd
Rtx
Fig. 6. IVG protocol relay selection after an accident.
The Reliable Geographical Routing in Vehicular Ad-hoc
Networks has been studied by Khil et al [9], where they
presented a RObust VEhicular Routing (ROVER) protocol.
They addressed the problem of broadcasting with floodingbased geocast algorithms that depends upon multicast
communications with end-to-end Quality of Service (QoS).
Like AODV, ROVER also floods control messages in the ZOR
while the data packets are unicasted. In this protocol each
vehicle is identified by a unique Vehicle Identification Number
(VIN) and the ZOR is of rectangular shape with their corner
coordinates specified. ZOF includes the source and the ZOR.
ROVER‟s packet format is a triplet [A, M, Z] where A, M
and Z corresponds to Application, Message and ZOR. The
message, M is meant for all the vehicles within a Z, ZOR from
an Application, A. Multicast tree is built from the source
vehicle to all other vehicles that lie inside ZOR by route
determination operation which begins upon receipt of a
message by network layer pushed by application layer. ROVER
is very much suitable for the scenario where it needs end-toend QoS.
Maihöfer et al. [10] came up with the idea of Abiding
Geocast which is a time stable geocast, requiring the
transmission of messages to all vehicles within a ZOR during
the geocast lifetime. Fig. 7, gives the building block of the
design space of Abiding Geocast. These building blocks can be
mingled in three different ways to achieve the Abiding Geocast.
First one is the Server Approach in which storage is done by
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16385
DOI: 10.18535/ijecs/v5i5.12
server and message hand over is not required at all. In the
second approach a node is elected in the ZOR to perform
message storage requiring the handover of messages before this
elected node leaves the ZOR. In the third approach each node
performs peer-to-peer task of storage and keeping the neighbor
information with handover of messages to the new node
entering the region.
Fig. 7. Building blocks of the design space of abiding geocast.
Celes et al. [11] came up with a geocast routing, GeoSPIN,
considering the data mining methods on individual vehicle‟s
daily trajectories records acquired through GPS. The
trajectories information is the spatial information which is used
in conjunction of sort-carry-and-forward method explored with
the opportunistic contacts of vehicles for message
dissemination. The GeoSPIN approach is split in two steps. In
the first step i.e., clustering of trajectories, data mining is done
on the daily trajectories of the vehicle‟s movement for
calculating the likelihood of the moving vehicle to take a
particular route. In the second step i.e., message forwarding,
after each node has a trajectory pattern, calculated in first step,
GeoSPIN disseminate message on the Encounter (n, r) and
Convergence (n, r) assumptions.
Rahbar et al. [12] came up with a Dynamic Time-Stable
Geocast Routing (DTSG) protocol aiming to keep a message
persistent within a geographical area for specified time interval
which can be scaled up, scaled down and even aborted, hence
the authors claimed the protocol as dynamic. The simulation
results of protocol showed it‟s free-wheeling with the vehicles
density, speed and wireless range. The message can be
contained over a particular geographic area for some time. In
protocol description authors coined four types of vehicles i.e.,
source, intended, helping and leader vehicles respectively. The
two stages of the protocol are pre-stable period in which the
region is populated with the messages and stable period in
which the protocol gets stabilizes within that region. The
DTSG protocol works well in sparse networks as it is shown by
the simulation results.
Constrained Geocast to support Cooperative Adaptive
Cruise Control (CACC) Merging was proposed by Wolterink et
al. [13] based on the probable vehicle‟s position in immediate
future time instead of the vehicle‟s instantaneous position. This
protocol adapts well with the increase in traffic but has
performance issues with other metrics. The application of this
protocol is to guide the joining of new vehicles inside a
currently moving linear batch of vehicles i.e., array of vehicles.
Chen at al. [14] came up with a Mobicast routing protocol
which is suitable for the application that needs space-time
ordination in vehicular networks. Vehicles in some geographic
zone i.e., ZORt, receives packet at time t, disseminated by a
source. ZORt is defined as a time function under certain time-
interval. As sometime vehicles miss to receive the packet
because of high velocity/mobility which is termed as temporal
network fragmentation and can be taken up by accurately
estimating dynamic forwarding zone. Three different but
related zones were coined as ZORt (Zone of Relevance at time,
t), Fig. 8, ZOFt (Zone of Forwarding at time, t), and ZOAt
(Zone of Approaching at time, t). ZORt defines a region with
event vehicle in the center such that all vehicles that are nearby
be able to collect the mobicast packet from event vehicle
successfully. ZOFt defines a region in which each vehicle
within the ZORt forwards the mobicast message collected by
event vehicle. ZOAt defines a zone for forwarding close to a
destination vehicle which forwards packets collected from
event vehicle. The spatio-temporal mobicast protocol is divided
into three steps of ZORt creation step, then mobicast message
delivery step and finally the ZOAt growing phase. The first step
identify the ZORt as a function of time, the second step
continuously disseminate control packets and the third step
solves the problem of temporal network fragmentation.
Fig. 8. Mobicast Zone of Relevance at any time instant
b) Multicast:
Conventional multicast routing algorithms are not applicable
for the VANETs as they were devised for the physically
connected wired networks. Many MANETs multicast protocols
are also suitable for VANETs since both are wireless but in the
latter case there is consideration of high mobility, frequent
topological changes etc. In multicasting a single node identifies
a group of nodes for information dissemination by multihop
communication. Based on the routing structure of the involved
nodes multicast routing algorithms gets broadly divided into
tree and mesh based routing algorithms. In the tree based
protocols claim packets are flooded by the host using the
optimal flooding techniques and involved nodes responses back
to the host along the backward path to form a multicast tree
rooted at the host. In mesh-based approach a mesh is sustained
consisting of connected part of the network that takes in all
recipients in a group. Senders and receivers constitute
multicast-group members and together with the forwarding
nodes (relay nodes) termed as tree or mesh nodes. Fig. 9, gives
the anatomy of multicast routing protocols.
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16386
DOI: 10.18535/ijecs/v5i5.12
needed for forwarding. With the increase in mobility within the
network ADMR switches to flooding mode and after short time
it settles again on the multicast mode. Packets are transmitted
using MAC-layer shortest delay path via the multicast
forwarding state. So in short ADMR does not use networkwide floods, adapts its behaviour and the high mobility can be
detected without the assistance of GPS.
Another tree based multicast protocol named MAV-AODV
will be discussed in next section.
Fig. 9. Mobicast Zone of Relevance at any time instant
In MAODV the AODV protocol is extended for
multicasting. For multicast AODV has to maintain a table i.e.,
multicast route table, having the fields such as group sequence,
next hop, group leader identification number, multicast group
identification number, hop count, lifetime etc. MAODV uses
broadcast technique for determination of routes. As depicted in
fig. 10 (a), when a mobile node either desires to attach in a
multicast group or needs to transmit data and it does not have a
routing path to that group, then it delivers a Route Request
(RREQ) message. The message is retransmitted by the nodes
till it gets received by a mobile node which is a member of the
multicast group tree which in turn sends the Request Reply
(RREP) message via unicast as shown in fig. 10 (b). The relay
nodes also mark the address of the mobile nodes from which
they collected RREQ packet in their routing table so that they
can make a backward route to the originator node of RREQ
packet. In case the originator node may receive multiple RREP
packets then the shortest route on the basis of hop-count metric
will be selected to forward the Multicast Activation (MACT)
packet, as shown in fig. 10 (c). After this exchange of message
the vehicle turn into a multicast group member and every
vehicle along the selected path from this vehicle to the vehicle
that recieves the MACT message becomes a forwarding vehicle.
The connectivity condition of the formed tree is monitored by
the first node to request membership, i.e., it becomes the group
leader. Link breakage eliminates the forwarding node from the
tree and the tree is repaired by re-establishing branches and
reconnection to the base tree is initiated. MAODV have simple
implementation but has many disvantages in VANETs scenario
i.e., long delays, overheads, low delivery ratio etc.
Adaptive Demand-driven Multicast Routing was proposed
by Jetcheva and Johnson aiming to reduce any non-on-demand
behaviour within the portions of on-demand protocols. In
ADMR protocol, for every source-destination pair a sourcebased frowarding trees are generated. Multicast sourceapplications are monitored for link breakage in trees and to
monitor the sources that have become inactiv. Two approaches
for repairing the link breakage are followed i.e., local repair
approach and global repair if former gets failed. For inactive
members the state is quietly ceased without any message.For
temporarily inactive senders the ADMR sends keep-alive
packets with increasing inter-packet intervals but when the
source becomes permanently inactive then the entire tree is
terminated. No control messages for tree maintainence is
required because the nodes are able to guess on the basis of
inter-packet time, the arrival of next multicast packet. Also
individual paths within a tree can be cut back when they are not
Fig. 10.
MAODV join operation
The proposal of Multicast Optimized Link State Routing
(MOLSR) is based on OLSR. The MOLSR is benefited by
gathering information of topology collected by OLSR protocol
which utilizes its topology control packets to form multicast
tree. For any source vehicle of multicast group, a multicast tree
is maintained in a distributed mode i.e., without any central
management. The multicast tree implements the shortest direct
paths from source to the members of multicast tree and on
detection of topology change the tree gets updated. The
overview of the protocol can be divided into three parts viz.,
tree building as shown in fig. 11, tree maintenance and tree
detachment. MOLSR protocol is classified under source-tree
based protocols. During tree building phase MC_CLAIM
packets are broadcasted to the entire network by the multicast
routers and this is done periodically. Source announces its
presence, whenever it requires to send message to a specific
group, by sending a SOURCE_CLAIM packet which lets the
members be connected to the tree. Optimized flooding of
OLSR is used for the flooding of messages in the network.
Upon receiving the SOURCE_CLAIM message by a member
which is not the part of tree, it explores in its table of multicast
routing to find the subsequent hop to reach the source and
making the subsequent hop its parent in the multicast tree by
delivering it the CONFIRM_PARENT packet. The tree
maintenance
is
done
by
SOURCE_CLAIM
and
CONFIRM_PARENT packets by OLSR methodology. For tree
detachment a node (leaf node) it dispatches a LEAVE packet to
its parent.
On Demand Multicast Routing Protocol (ODMRP) uses
forwarding group concept that comes under mesh-based
multicast protocol i.e., a subset of nodes forwards the multicast
packets through scoped flooding. Similar to reactive unicast
protocols, the ODMRP can be split in two steps i.e., requestphase and reply-phase. In ODMRP, group membership,
establishment and updating of multicast group is done by the
source. When a source vehicle wants to deliver packet, it floods
the JOIN_QUERY packet with data piggybacking and these
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16387
DOI: 10.18535/ijecs/v5i5.12
packets are regularly advertised for the route updating and
membership information. Upon interception of JOIN_QUERY
packet by non-member, initially it is checked for duplicity and
if it is not the case then upstream node ID is stored and then it
rebroadcast the message. Upon receiving the JOIN_QUERY
message by the multicast receiver it broadcast JOIN_REPLY
message. Upon receive of JOIN_REPLY message by a vehicle,
protocol checks for if the next vehicle ID agrees with its ID and
if it is the case then it authorizes itself by setting a flag to
become a forwarding vehicle and floods in the network its
JOIN_REPLY packet. Fig. 12, shows the membership
structuring and conservation with the forwarding concept. For
the nodes that desires to abandon the group the protocol quietly
ceases dispatching of JOIN_QUERY messages and the route
gets dropped upon not being refreshed i.e., a soft-state
approach. Main advantages of ODMRP is low bandwidth usage
and adaptability with topology changes
(a) source_claim Flooding.(b) confirm_parent msg.(c) source
multicast data.
-source
- multicast tree participant - client group
Fig. 11.
MOLSR tree building process.
(a) Join_query and Join_reply
the receiving of Join Query message by a multicast member it
waits for an interval hoping to receive more such messages to
select the best Join Query among the multiple Join Queries
available. Join Reply message are unicasted back to the
neighbor accordingly after the interval. During the data
forwarding step source dispatches messages to the forwarding
nodes which checks its flag to see whether it is in the
forwarding group or not for that multicast session and if it is
the case then it broadcasts the received packet. As compared to
ODMRP this protocol has less extra overheads.
c) Broadcast Routing Protocols:
Broadcasting refers to the dispatching of messages to all
nodes within the broadcast domain. This technique is
acknowledged as the most applicable technique for sharing
information about traffic, climate, emergency, accidents and
announcements. Common technique used in broadcasting is
flooding which leads to broadcast storm problem and
redundant message retransmission resulting in channel
congestion and decrease in reliability. Selective flooding
eliminates the redundant message retransmission as it lets only
selected relay nodes to perform retransmission of messages.
Each node has the responsibility of identifying the duplicate
packets to be discarded. Unicast protocols also use
broadcasting approach in their route discovery phase when
source don‟t have direct transmission range to the sender. All
the major broadcasting protocols proposed by researchers are
given in the Fig. 13, with their category but the discussion of
each individual protocols is beyond the scope of this work,
however we will discuss some protocols of each category.
(b) Forwarding group concept
Fig. 12. ODMRP join operation, membership
maintenance and group forwarding concept.
setup,
Destination-driven On Demand Multicast Routing Protocol
(D-ODMRP) is based on ODMRP which aims to enhance the
efficiency of multicast forwarding. In this protocol the
destination path is tendentious with those paths which traverses
another multicast destination and among them the path with
less cost is selected. The protocol implementation can be split
into three steps viz., join query process, join reply process and
data forwarding process. Here, all processes repeats at regular
intervals of time. In this protocol the Join Query phase is
somewhat modified from the ODMRP by adding
supplementary deferring-time at every mobile node which
receives the message. The deferring-time is calculated upon
how distant this received Join Query has left from the last
contacted group member and is proportional to the distance.
The deferring-time allows the Join Query to propagate quickly
through the less costly routes. During Join Reply phase, upon
Fig. 13.
Broadcasting routing protocol taxonomy
In table-based approach of broadcasting, each mobile node
holds directory of neighbors that is regularly updated by the
query and reply processes. The cluster-based approach of
broadcasting scheme splits the road topology in many clusters
and choose a cluster leader among the nodes forming the
cluster and then it exclusively performs broadcasting. Topology
based broadcasting protocols use network information, for
example density of nodes and connectivity of links to perform
broadcasting. Based on geographic areas messages are
disseminated in case of location based broadcast. In location
based approach each sending nodes adds its location which is
used by the receiving nodes. Distance based methods considers
the neighbor‟s relative distance and hop counts between source
and destination to decide whether to rebroadcast or not. Two
phases are there, first one is estimation phase and the second
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16388
DOI: 10.18535/ijecs/v5i5.12
one is broadcast phase. The probability based broadcasting
protocols assigns a predefined fixed probability to reduce
collisions and re broadcast by adopting persistence schemes.
The probability based approaches gives good results in dense
networks but has trifle significance in case of sparse networks.
Yu et al. [15] presents a Least Common Neighbor (LCN)
based table driven selective flooding protocols for
disseminating emergency messages for vehicular safety
applications. LCN method decreases the number of relay
vehicles that are in the same wireless range. Every sender‟s
message consists of its own neighbors directory and the
receiving vehicles matches its own neighbors directory against
the received message to determine whether it has least common
neighbors to be selected as relay nodes. If the common
neighbors are mostly same then the receiving node does not
broadcast the packet. Sun et al. proposed GPS-based Message
Broadcast for Adaptive Inter-vehicle communication [16].
They used the term TRAcking DEtection (TRADE) protocol
for their approach to organize the neighboring vehicles into
distinct categories to choose less number of vehicles for rebroadcast. Vehicles are put into three different groups namely
same_road_ahead, same_road_behind and different_road.
Then protocol selects the farthest vehicles (border vehicles)
from same_road_ahead and same road behind group. Every
vehicles are selected from different_road group. The sender
transmits the message with the border vehicles ID‟s and all the
receivers then decide whether to re-broadcast the message or
not by comparing the ID‟s within the message and its own ID.
Vengi et al. proposed Selective Reliable Broadcast (SRB)
[17] protocol for safety applications in VANETs. The protocol
proposes to minimize the broadcast storm complication for the
congested traffic scenarios where packet collisions occurs. The
whole vehicular network is partitioned into clusters with one
node among them elected as cluster head. Vehicles within a
cluster are independent of other clusters and they can‟t
communicate directly but via cluster heads as shown in Fig. 14.
A sender only forwards the messages to the cluster heads.
Arrival angle of the Clear-to-Broadcast (CTB) packet is
measured for the detection of cluster, which enables the source
vehicles to estimate the distances which if less than the
predefined threshold value then the vehicles are considered in
the same cluster. Within a cluster farthest vehicle is elected as a
cluster head and the whole process of cluster detection with
electing head is automatic and here the algorithm outperforms
the traditional approaches of broadcasting.
Fig. 14. Separated clusters of vehicles because of space
among clusters
Durresi et al. [18] presents a protocol called
BROADCOMM which aims to improve the quality of
broadcast in IVC with low network load maintenance. The
whole vehicular network is partitioned into virtual moving cells
and these cells move as the vehicles move. There are two levels
of categorical grouping in the vehicles viz., first level
comprises of all the nodes in a cell and second level is
represented by few geographically centered nodes with sensors
installed on it. Cell nodes can have inter cell communication
but with which they are in radio communication range. Second
level grouping‟s communication takes place when sensor
installed nodes communicates with nodes within the cell. The
hierarchical structure gives the protocol a choice of
differentiated service and good QoS and here is the advantage
of BROADCOMM when compared to the traditional
broadcasting protocols.
Tonguz et al. [19] proposed a protocol by the name DVCAST which is a distributed broadcast protocol for VANETs.
The protocol works in multi-hop broadcast fashion to work in
regular, sparse as well as dense traffics. The protocol is pure ad
hoc in nature with no infrastructure support and each vehicle
has GPS. The communicating device periodically sends out
hello messages at a frequency of 1 Hz. The per-hop routing
with local connectivity information assures the maximum
reachability of broadcast packets. Some routing parameters are
defined in the protocols viz., DFlg - for determining whether
intended recipient is moving in the same direction as the
source, Message Direction Connectivity (MDC) - for
determining whether it is the last vehicle in the group (cluster)
and Opposite Direction Connectivity (ODC) - for determining
if it is in connection with at least one vehicle moving in the
reverse direction. For vehicles which has DFlg set to 1 ignore
duplicate packets and if it has DFlg set to 0, then vehicle act as
a relay node and should do routing (per hop routing). The steps
followed by the protocol for appropriate manipulation of
broadcast packets depends on the density of vehicles (network).
Based on the work of DV-CAST Viriyasitavat et al. [20]
proposed Urban Vehicular BroadCAST protocol (UV-CAST)
which was acknowledged as the first work in broadcasting
routing protocols for urban scenarios. UV-CAST eliminates
broadcast storm and fragmented network complexities in
downtown sides to a large extent. The work was evaluated
against the metrics of network reachability, network overhead
and received distance, simulated in Manhattan mobility model
and real city of Pittsburg. Other topology based broadcasting
protocols are Vehicle Density-based Forwarding (VDF) [21]
and Density-aware reliable broadcasting in vehicular ad hoc
networks (DECA) [22].
Urban Multi-hop Broadcast (UMB) proposed in [23] aims to
eliminate the issues of hidden nodes, reliability and broadcast
storm in urban scenarios. The protocol is divided into two steps
viz., the first one is directional broadcast and the second one is
intersection broadcast. In directional broadcast sender vehicle
sends the packet to the far off node in the broadcast direction
and this does not require any topological information. In
second phase of intersection broadcast the repeaters installed at
intersection point has the responsibility to disseminate the
packet in all directions. The UMB protocol works without local
information‟s message exchanges thus reduces the overhead of
network. Intersection broadcast handling is shown in Fig. 15,
where node A reaches node B via directional broadcast as it is
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16389
DOI: 10.18535/ijecs/v5i5.12
out of the communication radio radius of repeater C installed at
intersection. As B is in communication radio radius of C, so it
can communicate with C and upon receiving of message by C
from B, C initiates the directional broadcast in the south and
north directions. D being in the transmission range of C also
receives the packet in east direction. Their further work
presents an Ad-hoc Multi-hop Broadcast (AMB) [24] which is
an ad-hoc extension of the UMB protocol. It does not require
the repeaters that are the greatest drawback of the UMB
protocol. When there is a street junction in the message
dissemination path, the vehicle closest to that junction performs
a fresh directional broadcasts to all road segments through a
fully ad-hoc algorithm.
Fig. 15.
UMB protocol‟s intersection handling [51].
Choi et al. [25] proposed Adaptive Location Division
Multiple Access (A-LDMA) protocol aiming to disengage
beacon traffic from the broadcast storm to accomplish more
persistent reliability of safety messages. Access of medium to
the vehicles is based on their location-to-time mapping and
their geographic location. The protocol reduces the contending
transmitters based on a TDMA schedule which is a simple
MAC level algorithm since it uses location-based deterministic
slot allocation approach.
Akkhara et al. [26] proposed Multi-Channel Cut-Through
Rebroadcasting (CTR) protocol for safety packets transmission
with the aim of minimizing the rebroadcasting vehicles and
overlapped rebroadcasting using multiple channels of the
available bandwidth. They gave the idea of giving preference to
the far off vehicle in the communication radio radius from
source vehicle to rebroadcast the message. Vehicles have two
transceivers installed and distinct channels are allotted to
vehicles for different hops to avert the collision during
broadcast. This protocol proposes the utilization of multiple
channels available from the total bandwidth.
Sun et al. [27] proposed broadcasting algorithm called
ODAM-C based on Optimized Dissemination of Alarm
Messages (ODAM) [28] aiming to improve the message
delivery proportion. The protocol employs two approaches
based on the forwarding aspects of ODAM viz., distance-based
approach for reducing the probability of losing packet by
calculating the angles between sources, forwarding and
receiving vehicles and redundancy approach for improving
packet delivery proportion.
Alshaer et al. proposed a probability and restricted zone
based broadcast scheme called as Optimistic Adaptive
Probabilistic Broadcast (OAPB) [29] to eliminate the broadcast
storm complication. The motive of their algorithm is to reduce
forwarding or rebroadcasting set of nodes to an optimal choice.
Each vehicles rebroadcast probability adaptively changes
within two hops depending on its local information. Periodic
hello messages enables the nodes to get local information for
the estimation of local vehicles density based on which nodes
dynamically calculates its rebroadcast probability. Vehicles
with larger probability value are assigned a shorter delay time
to rebroadcast.
Reception Estimation Alarm Routing (REAR) [30]
guesses the reception possibilities of alarm packets for the
moving vehicles. Instead of selecting those nodes as relays
which are far off this protocol gives preferences to those
vehicles which has highest probability to relay packet based on
real wireless channel. Periodical beacons are used to collect
information regarding location and size among the neighbors to
maintain neighborhood list. The alarm message also contains
the neighborhood directory table and direction of message
propagation. Those vehicles which received the alarm packet
and are moving in the same direction of alarm source vehicle
can participate in relay node election. Calculation of contention
lag based on receipt possibilities of neighbors leads to
contention phase. The time taken for relay of packet by the
node is proportional to the contention delay and when the
vehicle is trying to relay packet to other vehicle and hears the
alarm packet, it cancels its contention phase and hence
redundant broadcasting is avoided in this way.
B. Based on Prerequisite Knowledge
1) Topology-Based Routing:
Fig. 16 shows the taxonomy of topology based routing which
can be categorized as proactive (table-driven), reactive (ondemand) and hybrid routing protocols. Proactive algorithms
apply the concept of shortest path algorithm for unicasting the
packets. Neighboring nodes information are stored in tabular
form which gets shared between the vehicles for the updation
of the network‟s topological changes. On-demand or reactive
algorithms catches routes on demand by flooding network with
route request messages which sometimes leads to network
clogging and there is delays in route discovery but is suitable
for VANETs as the topology changes very frequently with
time. In hybrid protocols the aim is to combine the goods of
both proactive and reactive algorithms. The routing is
originally settled by proactively discovered routes then after
routing is served using reactive algorithms.
DSDV is one of the earliest ad hoc routing protocol adapted
from MANETs and is based on Bellman-Ford algorithm.
DSDV guarantees the loop free paths and decreases the
convergence time compared to its earlier protocols. The
protocol maintains a table for each node where each table
contains information about all accessible network nodes with
the count of hops to reach them and each table entry is marked
with sequence counters by the destination nodes. The sequence
number is even if the link is ok otherwise it is odd. Consistency
maintenance of routing tables is maintained by regular
broadcast of routing tables to the neighbor nodes when there is
a change in topology or new information is available.
Global State Routing Protocol (GSRP) proposed in [31],
maintains a global knowledge of the network topology by
exchanging vectors of link states between neighbors during
routing information exchange. Initially every node have
unfilled table of neighbor nodes and unfilled topology table but
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16390
DOI: 10.18535/ijecs/v5i5.12
they pick ups information about their neighbours by scanning
the sender field of each message in its inbound queue. In
Fisheye State Routing (FSR) [32] the GSRP protocol is
improved. In FSR each node has unique ID and maintians three
tables viz., next hop table, distance table and topology table. It
also have a list of neighbour. Neighbor list contains adjacent
node IDs and topology table has two parts for each destination
which indicates the link state condition noted by the destination
and the time stamp indicationg the time destination node has
noted that link state condition. The distance table provides the
shortest path between pair of nodes. Adjacent nodes are
frequently updated with respect to the further nodes and the
updated messages do not hold information about all nodes
hence utilizes bandwidth properly.
unicasting and multicasting. In case a source needs to transfer a
message to a node and has no routing clue then the route
discovery process is initiated by broadcasting RREQ packets
which is rebroadcasted by its neighbors until it reaches the
destination as shown in fig. 17. When RREQ packet reaches
destination it responds back with RREP beacon through the
path which the protocol learned by backward learning.
(a) Propagation of RREQ
(b) RREP path to the source
Fig. 17. AODV route discovery process.
Fig. 16.
Topology-based unicast routing taxonomy
OLSR is the modification of old LSR protocol cutom-fit to
the wireless requirements. The idea is to select multi pointrelays (MRPs) by each node among its neighbor nodes for
minimizing the packet streams overhead in the same area. The
neighbor which are not selected in the MRP set only receive
the message but do not retransmit it again. Regular hello
messages are transmitted for link information collection. The
criteria for MRP selection is simple in the way that the packets
retransmitted by these nodes should be received by all nodes
which are two hops away from the sender.
TORA [33] is a distributed routing protocol where route
optimality is sacrificed in favor of lower numbers of overhead
messages. The protocol‟s implementation is separated in three
basic functions of route creation, maintenance and
elimination.The modelling of network is done as a set of finite
nodes and undirected links. Due to the nodes mobility the set of
links changes with respect to time. Each initially undirected
link may subsequently changes to undirected and directed link
from one node to other and vice versa. For route creation
undirected links are changed to directed one when sender has
no route leading to destination with the help of query-reply
messages and a Directed Acyclic Gaph (DAG) is constructed
rooted at the destination. TORA reacts to maintain routes for
any topological changes so that the routes leading to
destination can be established within a limited time interval.
When network partitioning isolates the destination node then
the directed link are changed to undirected. This protocol
ensures loop free routing as the packets are dropped by the
neighbors of the sender if it has no downward link to the
destination.
AODV and DSR are other on-demand protocols adapted
from MANET. In AODV some nodes initiates a route
discovery process to reach destination only when needed.
Counting-to-infinity complication of other distance-vector
algorithms is avoided in AODV by adopting the DSDV
concept of sequence numbers. AODV can be used for both
DSR is a reactive algorithm sharing similarities with AODV
as it also forms routes on basis of demand by sender. The
sender maintains the whole route within the header of the
packet meant for destination node. Retransmission of packets
by intermediate nodes is based on the route captured in the
header. In case sender don‟t have any information about the
route to destination then process of route discovery is initiated
and sender broadcast the route request packet to the
neighboring nodes to be broadcasted again till it is captured by
the destination. Sender host receives a route reply message if
the route discovery process is successful with the listing of
network hops sequences. For maintaining routes no explicit
packets is transmitted and in case of broken route the source
can attempt with other known route or again initiate the route
discovery phase.
Zone Routing Protocol (ZRP) [34] is based on the idea of
routing zone which is a group of nodes whose radius is referred
as the zone radius. Routing zone is defined for each node and it
has to know the topology of only its own zone and they get
updates of topological changes of their corresponding zone
only. So a large network is partitioned in zones and the updates
are broadcasted locally. There is an intra-zone communication
between the nodes but if a sender from one zone wants to send
information to the other zone‟s node then it sends the query
messages to its border nodes which again retransmits the
message to its border nodes until it reaches the destination or
the hop-count reaches zero.
Hybrid Ad hoc Routing Protocol (HARP) [35] was proposed
by Nikaein et al. which combines some features of both
reactive as well as proactive protocols. In this protocol inter as
well as intra zone routing is performed which depends on
whether the communication to be performed is within the same
zone or outside the zone. The inter zone communication is
reactive in nature while the intra zone communication are
proactive. Distributed Dynamic Routing (DDR) algorithm is
used for zone creation which is a logical structure with respect
to the network properties.
2) Position-Based Routing:
These routing algorithms relies on geographic position
information of all nodes and their neighboring nodes using
GPS devices. The geographic position information is used for
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16391
DOI: 10.18535/ijecs/v5i5.12
routing decision and don‟t require to manage any routing table.
They are further split upon as non-delay tolerant network (nonDTN) and delay tolerant network (DTN) which we will discuss
in latter section.
Fig. 18.
control with bounded delay for all users. The area where
vehicles are located are divided into smaller areas with one-toone mapping between the bandwidth divisions using any
TDMA, CDMA or FDMA technique. SDMA is self-starting
and self-maintaining protocol.
Greedy forwarding approach of GPSR.
Greedy Perimeter Stateless Routing (GPSR) was proposed
by Karp et al. in which the forwarding of packets is based on
the routers and destination‟s position. GPSR uses two different
algorithm for packet forwarding namely greedy forwarding and
perimeter forwarding. In greedy forwarding the sender sends
the packet to the neighbor which is closest to the destination
because the packets are marked with the destination‟s position.
Thus a greedy choice is employed in every hop for forwarding
the packet until it reaches the destination. Fig 18 shows the
greedy forwarding approach. The positions of all nodes are
determined by the simple hello messages containing sender‟s
ID and position. Nodes are purged from the table if for a long
time no hello messages is received from them. Perimeter
forwarding is done when greedy forwarding fails in case when
the sender is physically closer to the destination than its
neighbor but is not in the direct range with the destination i.e.,
a problem of local maximum. In this case the graph with
sender, its neighbors and destination is traversed by right hand
rule.
Greedy Perimeter Coordinator Routing (GPCR) [5] is based
on GPSR but alleviates the problem posed by the obstacles
which causes network partitioning. Like GPSR it also has two
different algorithms with the same aim but its greedy
forwarding scheme is somewhat restricted and a recovery
algorithm when it‟s greedy forwarding algorithm fails. The
algorithm uses road‟s junction point as vertices and streets as
edges to construct the planner graph without any map support
so no algorithms is required for the graph construction. When
there is no local maximum problem the algorithm uses its
greedy forwarding approach. The packets are routed only along
the streets and the decisions are made at junction points of
streets and this alleviates the problem of blockage by buildings
etc. Fig. 19, shows the contrast between the normal greedy
approach and restricted greedy approach in which if regular
mode is used then the packet has to follow the path, S -> 1a ->
1b -> 2a -> 2b -> D but in case of restricted greedy approach
the packet will follow the path, S -> 2a -> 2b -> D. The nodes
which are at the junction are called coordinators, which are
chosen randomly by the sender or forwarder, and each
coordinator broadcast their position‟s information. The
recovery algorithm do greedy routing to the next junction point
where the decision of which route to be taken by the packet is
made. Thus the decision making is done by coordinators nodes
only which are located near junction point.
Space Division Multiple Access (SDMA) approach was
proposed by Bana et al. which allows the medium access
Fig. 19. Greedy forwarding approach verses restricted greedy
forwarding.
Connectivity Aware Routing (CAR) [36] was proposed by
Noumov et al. which locates destination and finds connected
paths between source and destinations. The CAR protocol has
four parts viz., destination location with path discovery, data
forwarding along discovered path, maintenance of paths and
recovery for broken links. Beacons are used for direction and
speed information and caching of successful routes between
various pairs of source and destination is also used.
Edge-node Based Greedy Routing (EBGR) [37] was
proposed by Prasanth et al. in which the edge nodes of
transmission range are selected as a next hop node for
forwarding data with considering the nodes which are moving
in the same direction as the destination node. So the protocol
has three parts viz., neighbor node identification for collecting
all direct neighbor information, node moving direction
identification for identifying the nodes which are moving in the
destination‟s direction and edge node selection for selecting
next forwarder of data packet. Contention Based Forwarding
(CBF) [38] proposed by Füßler et al. is a greedy position based
beaconless protocol which exploits the trail method for nexthop selection. The sender sends its packet to all its direct
neighbor and let them select the next forwarder by using the
proposed distributed timer-based contention process. Each
CBF packet has ID and position of its sender and final
destination. Upon receiving of CBF packet by a neighbor
which if not a final destination a timer is set for forwarding the
packet based on the progress towards destination. Since no
beacons are used so it utilizes the bandwidth more efficiently.
We will discuss Intersection based Geographical Routing
Protocol (IGRP) which is a position based unicasting algorithm
in latter part of this section under bio-inspired routing
algorithm. Position based routing are further classified based
on delay sensitivity and toleration which is also discussed
further in this section.
3) Map-Based Routing:
Multi Hop Routing Protocol (MURU) [39] presented by Mo
et al. introduces new reliability metric called as Expected
Disconnection Degree (EDD) for finding robust path for urban
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16392
DOI: 10.18535/ijecs/v5i5.12
VANETs scenarios. EDD uses the factors such as position,
speed and trajectory for route quality estimation. MURU is
pure ad hoc in nature without any infrastructure support. The
main aim of MURU is to eliminate the problems caused by
buildings and other wifi interferences. Each vehicle knows its
position using GPS and uses external static roadmap. Shortest
trajectory is calculated using roadmap by the sender to its
intended destination.
Geographic Source Routing (GSR) [40] proposed by
Lochert et al. was the first protocol which was evaluated over a
realistic vehicles trajectory pattern which aims to alleviates the
problems faced by the general position based routing protocols
because of radio obstacles. The protocol is ad hoc in nature and
uses city map. A Reactive Loaction Service (RLS) is used for
finding the positions of nodes. The sender uses street map for
the knowledge of sequences of junctions the packet has to
travel to reach the destination and the packet header contains
these junction sequences. The sender optimally floods the
network with request packet to know the position of destination
which responds back by replying its position so that the GSR
protocol finds the route to destination using map.
Shortest path based Traffic light Aware Routing (STAR)
proposed by Chang et al. which considers the traffic lights and
traffic pattern for routing decisions. It considers the effect of
traffic lights on the mobility of vehicles with the prior
knowledge of road topology as vehicles gets stopped by the red
lights and move when there is green light. Before reaching the
street junctions packets are forwarded in normal greedy
approach. At the street junction STAR protocol checks for if
the destination is connected or not if not then the packet is
forwarded to the closest green light segment with the
destination.
Anchor based Street and Traffic Aware Routing (A-STAR)
proposed by Seet at al. which addresses the problem of local
maximum and also uses the route information of city buses for
finding anchor points with good connectivity for packet
dissemination. Like STAR protocol also takes traffic lights and
prior knowledge of road topology into account. The source
packet is marked with the anchor points through which the
packet has to travel in order to reach the destination and in
between greedy forwarding technique is used. The street maps
are statistically rated with the number of city buses that ply on a
particular route. The protocol inner core idea is based on the
already discussed GPCR protocol.
Road Based using Vehicular Traffic (RBVT) was proposed
by Nzouonta et al. with proactive and reactive versions of the
protocol. The RBVT protocol takes advantage of real time
vehicles trajectory information for creating road based path
with junctions providing connectivity between them. Each
vehicle in RBVT has GPS system, maps and navigation system
that gives position of node on road. In reactive version of
RBVT the path is made of road segments divided by junction
points with vehicles providing connectivity between junction
points. In proactive version of RBVT the periodical discovery
and dissemination of road-based network topology is done to
preserve steady state of network connectivity at each node.
4) Path-Based Routing:
Vehicle Assisted Data Delivery (VADD) proposed by Zhao
et al. is a path based routing protocol for VANET. The idea of
protocol is based on the fact of carry and forward until a new
node comes enough close to make the forwarding or delivery of
the packet. VADD is particularly suited for the sparse network.
Each node know its position and is equipped with the
statistically rated digital map. VADD uses wireless radio
channels to forward the packet but the nodes with higher
speeds are preferred with periodical path maintenance all over
till the packet is delivered. VADD synthesize three types of
packets viz., straightway_mode, intersection_mode and
destination_mode on the basis of nodes location carrying the
packet as shown in fig. 20. VADD delay model employs
stochastic model to guess delivery postponement.
Fig. 20.
Packet modes in VADD
C. Based on Delay Sensitivity and Toleration
1) DTNs:
DTNs find their applicability in disaster area, military
operation and emergency response networks. Distributed
Adaptive Routing (DAR) proposed by Khanna et al. which
aims to achieve high network connectivity with less network
transmissions. DAR uses gossip protocols pioneered by Xerox
PARC for DTNs to show the phase transition characteristic of
delivery ratio in DTNs. Adaptive gossip probability algorithm
uses phase transition value for gossip probability computation
for every node. The computed probabilities limits network
transmissions by letting each node to decide probabistically
whether it should rebroadcast the packet or not. We have
already discussed VADD in path based protocols which is also
a DTNs routing protocol.
Scalable Knowledge-based Routing (SKVR) Architecture
for public transport networks proposed by Kanere and Ahmed
is based on the analysis of vehicles trajectory trace files which
shows that the public transport has the characteristics of DTNs.
They have proposed the public transport arrival-departure
timings knowledge as an aid to routing protocols and
partitioned these knowledge in static and dynamic one. Static
knowledge corresponds to the fixed planned timings of public
transport whereas dynamic knowledge corresponds to the
variations in the static knowledge. The network of public
transport has hierarchical structure with inter-domain routing
among different public transport routes and intra-domain
routing within a particular public transport route.
Social-based Privacy-preserving forwardING (SPRING)
[41] is a routing protocol for DTNs which uses RSUs for
packet forwarding assistance. RSUs are used only when no
reliable next-hop vehicle is available, i.e., SPRING has V2I
communication. In this protocol social degree of each road
junctions/intersection points is introduced which is defined
heuristically for the placement of RSUs at the intersection
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16393
DOI: 10.18535/ijecs/v5i5.12
points which have high social degree. RSUs placed according
to the value of social degree of intersection point provides high
connectivity because maximum number of vehicles will cross
from these junctions and at the same time reduces the cost.
SPRING also takes authentication, privacy and attack
resistance into account.
Geographical Opportunistic (GeOpps) Routing protocol [42]
is a geographical with delay tolerancy approach which uses
opportunistic contact between vehicles for data forwarding.
Information from the navigation system is used for the choice
of next hop which takes radial separation from the final
destination of the packet as its basis. Every forwarding node
calculates its nearest point to destination using the navigation
system to decide whether to keep the packet or to forward it to
a neighbor on the basis of estimated minimum-time required by
the packet to be intercepted by the destination. The calculation
of minimum-time is done by using a utility function given by
the authors.
GeoSpray proposed by Soares et al. uses a hybrid approach
between single and multi copies store and forward approach
with node‟s geographhical position data. It asynchronously
spreads limited bundel copies which are IP datagrams packets
to be processed by bundle layer between network and MAC
layer. GeoSpray is inspired from GeoOpps as it also uses
navigation systems for the node‟s position information. Once
the multiple bundle copies are spread in the network the
protocol switches to single copy forwarding scheme. All nodes
when interacts with other node exchanges information for the
deletion of bundles that have already been delivered then they
check whether the packet they are holding has final destination
among them or not, if it is then those bundles are delivered and
purged from the buffer. In the next step they exchange
information about the bundles they are holding to determine the
best carrier node for each bundle stored with them. The
algorithm is symmetric as same is running between the
interacting nodes.
Spray and Wait protocol introduced by Spyropoulos et al., is
a tradeoff between full-fledged broadcasting and optimal
broadcasting approach with epidemic routing in combination.
This protocol can also be viewed as a adjustment in between
multi copies and single copy plans. In this protocol a number of
data packets are sprayed as the sender encounters other nodes
then the sender waits for some time. Binary Spray and Wait
technique is used which can be defined as “Source node
originally begins with „M‟ copies of message and any node „N‟
having n > 1 copies of messages if meets another node „X‟ that
have no copies then „N‟ hands over to X ہn/2 ۂkeeping ڿn/2ۀ
messages but when there is only a copy left, then it goes for the
direct transmission” In wait phase the author states that “If
spraying phase is not able to discover the destination then every
node having message copy goes for the direct transmission
(i.e., will forward the message only to its destination).”
D. Based on Bio Inspiration
Bio-inspired algorithms can be defined as those algorithms
which takes a more evolutionary approach to learning by taking
ideas from life science and tries to copycat the behavior of
natural breeds. The protocols which we discussed in the
previous subsections takes traditional computational approach
for routing solutions. These algorithms can further be divided
into swarm intelligence, genetic algorithms and evolutionary
algorithms. Swarm intelligence is the unified habits of
distributed but self-coordinated communities like ant
community, bird flocking, animal herding, fish schooling, etc.
It consists of a community of simple operating agents which
interacts restrictedly with one another and follows some rules.
Swarm intelligence pattern examples are particle swarm
optimization, ant community optimization (ACO), bee
community optimization (BCO), bat algorithm, river formation
dynamics, etc. Below we discuss some of the bio-inspired
algorithms.
A delay sensitive vehicular routing protocol using ACO
proposed by Li et al. [43] falls under DTNs routing protocol
which uses ACO techniques to choose the least delay path for
forwarding packets. This protocol uses streets/highways
junctions as anchor points i.e., RSUs for data delivery
assistance. ACO algorithm is used for alleviating the problem
of hard non-deterministic polynomial in routing. Initially
optimal paths are established between source and destination
using unicast or optimal broadcast transmissions with reactive
approach. For route maintenance and path extension the
protocol takes table-based proactive approach. GPS, navigation
system and maps are pre-installed on each vehicles so that the
vehicles can know their geographical positions when needed.
Links relaying quality is estimated based on the additive
relaying delays between two junction points. Reactive forward
ants of ACO concept is used to find possible paths from source
to the closest junction which generates the reactive backward
ants back to source. The reactive backward ant‟s packets marks
pheromone at each junction using the delay values. These
pheromone values are used to select the best junction‟s RSU
among the available to relay the packet with simple greedy or
carry forwards between the junctions along road segments. The
algorithm uses reactive, proactive and V2I approach for its
implementation.
Mobility-aware ant colony optimization routing called as,
MAR-DYMO was proposed by Correia et al. [44] also uses
ACO techniques for making routing decisions in urban
scenarios. Kinetic graph framework devised by Harri et al., is
used in this protocol for collecting information regarding
vehicle‟s position and speed. The routing table at each node
maintains the information of pheromone in it and represents the
route qualities which is high if the route request and reply
packets are both received at the same node. During packet
forwarding the pheromone levels of routes are considered to
select the best path between the source and destination. The
route discovery process of MAR-DYMO is reactive compared
to DYMO and uses kinetic graph framework.
Multicast with Ant Colony Optimization for Vanets based on
MAOVD (MAV-AODV) proposed by Souza et al., also uses
ACO for building optimized and stable multicast trees. In this
protocol beacon packets act as pheromone of real ants.
Neighbors within its radio range send beacon messages at
regular intervals among themselves to know positions and their
mobility in the form of position vector and velocity vector.
Link lifetime calculations is based on the exchanged position
vector and velocity vector. Route request packets are
broadcasted for multicast group member‟s discovery from the
source for the destination. Each route request packet stores the
route lifetime estimated earlier for making stable multicast
trees. Upon receiving of route request packet by a multicast
group member it replies to source node along reverse path
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16394
DOI: 10.18535/ijecs/v5i5.12
using unicasting by taking route‟s pheromone level and hop
count into consideration as shown in Fig. 21. The MAVAODV simulation results are better than DYMO discussed
earlier in this sub section and generates stable multicast tree
than other conventional multicast protocols such as MAODV.
Fig. 22.
Fig. 21.
Route reply example of MAV-AODV [39].
Intersection-based Geographical Routing Protocol (IGRP)
proposed by Saleet et al., is based on genetic algorithm in way
that the protocol makes selection of street junctions to forward
the packet to the internet gateway. The work of Saleet et al.
aims to provide effective and reliable communication. Every
vehicle in this protocol is assumed to have the GPS, digital
map and navigation system. The street junctions are considered
as vertices and the roads connecting these junctions are the
edges of the graph abstracted for the street map. In this
protocol the whole network is considered to be network among
mobile vehicles and fixed internet gateways installed at the
street junctions as shown in Fig. 22. The aim of this protocol is
to forward the packet from source to the nearest internet
gateway as soon as possible with taking QoS, bit error rate
(BER), connection probability etc., into account. The purpose
of the installed gateway is to supply the information about the
best route for forwarding the packet. Gateways has all the
information of local network topology made up of all the
nearby vehicles as all nearby vehicles gives updates of its
mobility to its nearby gateway. Gateways acting as the location
server helps in constructing the routes between itself and other
vehicles. Between two gateways the mobile vehicles will
transmit the packets but the routes between them are not stable
because of mobility, so IGRP takes the approach of backbone
routes consisting of street‟s junction point only. The gateway
selects that backbone which has highest probability of
connection and also sufficient traffic and sends this information
to the source so that these information should be added in the
packet header. The packet header‟s route information helps the
forwarders to geographically route the packet to the ultimate
destination. The authors of IGRP also gave equations for
calculating connectivity probability, bit error rate, delays, hop
counts and transmission range.
Message routing in IGRP [69].
E. Based on Accommodating Network
1) Homogeneous Network:
For V2V and V2I communications IEEE proposed a
standard by the name of DSRC which has WAVE as its core
part. DSRC is comprised of the set of IEEE 1609.x standards
built over the IEEE 802.11 standard as shown in fig. 2. When
all the vehicles uses same underlying wireless access
technologies then it is the case of homogeneous network.
Vehicles can communicate with the technologies like cellular,
WiFi, WiMAX, satellite, Bluetooth, LiFi and DSRC/WAVE.
However DSRC is best for vehicular communication as it is
specifically designed standard to meet the extremely short
latency requirement of vehicular environment. DSRC and
WAVE are sometimes used interchangeably.
Ho et al. [45] presents DSRC system implementation. Many
projects have been undertaken and some are going on to bring
homogeneity in vehicular environment for computing and
communication. Table I enlists objectives and references of the
projects undertaken as well as going on projects for bringing
standards and homogeneity among VANETs communication.
To alleviate the issues of heterogeneity, security, cost, etc.,
David Padi [46] proposed Vehicular Information and
Communication Technology (VICT) system to set out
technological solution for robust, resilient and secure
communication with managerial accountability in vehicular
environment. Hung et al. [47] used Worldwide Interoperability
for Microwave Access (WiMAX) for implementing video
based safety application. Particularly WiFi and WiMAX are
suitable for non-safety applications that require video
streaming as it gives high data transfer rate of up to 63 Mbps in
downlink and upto 28 Mbps for uplink.
Jaber et al. [48] proposed the combined usage of WiMAX
and DSRC for vehicular communication with the aim of
providing broadband internet access to DSRC enabled vehicles
being served in WiMAX region. Another technology that have
place in vehicular communication is Light-Fidelity (LiFi ) [49]
which uses visible light communication (VLC) technology to a
further extent. Bluetooth is also becoming popular for vehicular
communication but till now it has not been recommended as the
replacement for on-board communication interface of
DSRC/WAVE.
2) Heterogeneous Network:
When vehicles uses different radio access technologies then
it is the case of heterogeneous vehicular environment. The
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16395
DOI: 10.18535/ijecs/v5i5.12
main challenge of routing protocols in heterogeneous
environment is inter-communication understanding of different
underlying wireless access technologies. The simulation results
gives good performance when it is assumed that all vehicles
and associated infrastructures uses same technologies. Routing
in heterogeneous vehicular environment especially needs to
overcome the challenges of handoff, computational complexity,
transmission time, cost, coalition management, mobility
management and topology control problems. Vertical handoff
can be defined as a “mechanism of switching between the
categories of connections a mobile node uses to have access of
supporting infrastructure”. In simple language it can be stated
as the change from one type of network technology to the other
type is called vertical handoff. Nodes mobility is usually
supported by the mechanism of vertical handoff techniques.
Shafiee et al. [50] proposed an optimal distributed vertical
handoff for heterogeneous vehicular networks based on
assumption that vehicular network involves cellular system and
WLAN. Kim et al. [51] proposed fuzzy logic based handoff for
minimizing transfer time and cost. Their work proposes the
selection of best communication technologies available around
vehicle. Li et al. [52] proposed algorithms for topology control
in wireless network that can be effectively used for
heterogeneous vehicular environment also. They proposed
Directed Relative Neighborhood Graph (DRNG) and Directed
Local Minimum Spanning Tree (DLMST) algorithm both of
which are localized topology control procedure which lets
each node to select its neighbor and also tune its transmission
power. Another approach of cluster formation that withstand
the mobility pattern can also alleviate the problems inherent
with heterogeneous networks. Clustering algorithms can be
divided into mobility based and non-mobility based which
includes direction of movement, signal strength, transmission
range, probabilistic approach, etc., as cluster formation
parameters. A cluster head is selected by the nodes that
constitute a particular cluster which act as a gateway between
different network technologies. For connecting a source vehicle
to 4G LTE advanced infrastructure Zhiona et al., proposed a
clustering algorithm. The algorithm uses some information
collected by source viz., signal strength, load and connectivity
for choosing the gateway. Amalgamation of routing techniques
with different tradeoffs can also solve the problems faced by
heterogeneous vehicular network environment. One such
proposal were made by Shaifee et al. [50] in the paper titled
WLAN-WiMAX Double-Technology Routing (WWDTR).
3. Assessment Approaches for VANETs
As we mentioned earlier that evaluation of a routing protocol
in VANETs is necessary, indispensable and struggling task, so
in this section we bring some assessment methodologies into
picture. Mostly protocols are accessed using simulators
because conducting live real-time experiments are costly,
tedious and risky events. Once the protocol passes all the
simulation tests with expected results then it can be tested in
the real time vehicular environments. Before deploying any
protocol in real world we should have enough facts and figures
that it will attain the proposed aims without any risk. For
assessment we need network simulator and vehicular mobility
simulator with integration of these two.
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16396
DOI: 10.18535/ijecs/v5i5.12
number of vehicles, density of vehicles, speed change of
A. Vehicular Mobility Simulator
TABLE II
Network Simulator Comparison
Developme
nt
Language
Simulatio
n
Language
GNU
GPLv2
C++
C++/OTcl
GNU
GPLv2
C++
C++/Pyth
on
Simulato
r
License
ns-2
ns-3
Source
Code
Availabilit
y
Inclusion of
Mobility Model
Mobility
Model
Supported
Windows(Cygwin),
Linux, FreeBSD
Yes
No
VanetMobiS
im, SUMO
Windows(Cygwin),
Linux, FreeBSD
Yes
No
VanetMobiS
im, SUMO
Supported OS
C++
Windows, Linux
Yes
No
VanetMobiS
OMNeT Academi C++
++
c
im, SUMO
Vehicular mobility model is required to study the effects of vehicles, clustering of vehicles at junction points and moving
direction of vehicles on the routing protocols. The choice of
appropriate mobility model is very crucial as it determines the
accuracy and reliability of the simulation results. Fig. 23 gives
the classes of mobility models at a glance so far proposed by
researchers. In synthetic modelling mathematical exemplary
models like stochastic, traffic stream, car following, queue and
behavioral are used whereas in survey modelling whereas in
survey modelling extracted survey data are used for mobility
pattern generation. Trace based approach of modelling directly
uses real mobility traces generated by various sources. In traffic
simulator based approach traces of mobility pattern form traffic
simulator is used.
Simulation of Urban Mobility (SUMO) [53] is designed for
urban traffic simulation and is open source technology. With
macroscopic vehicular mobility it can simulate large number of
vehicles simultaneously plying under the given street topology
constraints. SUMO can be used both in command mode and
GUI mode. VanetMobiSim [54] is also a popular traffic
simulator published by university of Stuttgart and is a Java
based application supported on Windows, UNIX and Linux
platforms. VanetMobiSim is an enhancement to the
CanuMobiSim [55] and allows to define street topologies using
TIGER map and clustered Voronoi graph with input as XML
file. STRAW [56] is another mobility simulator developed as a
part of “Car-to-Car Cooperation” (C3) project and defines the
streets topology using real maps. STRAW also supports the
implementation of lane changing, traffic signs and traffic lights.
B. Network Simulator
Network simulator is required to study the effects of
communication range, interference, medium access and
topology changes. Among the available network simulators
OMNeT++, ns-2 and ns-3 are most popular discrete event
driven simulation tools. While ns-2 and ns-3 comes under GNU
GPLv2 licensing, OMNeT++ is for academic purpose. Table II
gives a comparison of these network simulators.
C. Integrated Simulation Environment
Traffic simulators generates what we call trace files and
introduction of these trace files to the network simulator is the
approach for VANETs application or routing protocols
simulation. On these basis we have two different kind of
approach i.e., tightly assimilated approach and loosely
assimilated approach. Both approaches are used for VANETs
simulation but the difference lies in the mobility generation.
Tightly assimilated approach assimilates both network
simulator and traffic simulator into one with latter having the
responsibility of vehicular mobility and former having the
responsibility of wireless communication. In loosely
assimilated approach the trace files independently generated
are introduced into the network simulator for mobility.
ITERIS and Veins are most popular tightly assimilated
framework for VANETs simulation and gets easily combined
with SUMO. Both ITERIS and Veins have two way
communication and while former uses ns-3 as network
simulator latter uses OMNeT++ as network simulator with full
featured WAVE model support. Both of these also gives GUI
based visualization.
4. Conclusion
VANETs research and projects started as an offshoot of
MANETs feeling the need of inter-communication between
vehicles for safety application, traffic warning and to eliminate
or minimize the accidental hazards. With the maturity of
technologies VANETs got their place in safety, non-safety,
commercial,
non-commercial,
entertainment,
platoon
management and parking system applications. Many
researchers, universities, government and non-government
organizations came forward for the realization of VANETs
applications. The core of all the application lies in the
exchange of data in such mobile and dynamic environment and
for this many routing protocols were contributed by the
researchers. Some routing protocols were successfully
implemented with the current technologies and some will
become feasible latter with the advancement of technologies
and standards. In this work we discussed the existing and
ongoing technologies which VANETs require such as WAVE,
DSRC, WiMAX and devices like GPS, OBU and RSU.
The aim of this work was not to produce a voluminous work
but to give an insight into the various classification strategies
one can go with. We have discussed major existing protocols
and many not because of limited space. Routing techniques
employ various graphical, probabilistic and hybrid approaches
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16397
DOI: 10.18535/ijecs/v5i5.12
for its realization and implementation. Classification taxonomy
were based on transmission strategies, knowledge needed,
delay awareness, accommodating network and motivating
inspiration.
Based on transmission approach, unicast, multicast, geocast
and broadcast routing protocols were included. Unicast routing
routes packets from single source to single destination while
multicast and geocast routes packet from source to multiple
destinations. Geocast uses the concept of ZOR and on the other
hand multicast may be tree or mesh based on the routing
structure of involved vehicles and its road networks.
Broadcasting protocols tries to eradicate the problems of
broadcast storm, redundant retransmission and channel
congestion.
Many routing protocols needs some prerequisite knowledge
about the vehicular environment for their realization like
topology-based, position-based, map-based and path-based
routing protocols. Further they are hierarchically classified as
proactive, reactive and hybrid routing protocols depending on
the methodologies they employ. Topology based routing
employ the concept of shortest path algorithms like Dijkstra‟s
and Bellman Ford‟s algorithm. Position based routing relies on
geographic position information of all nodes and their
neighboring nodes using GPS devices.
DTNs routing protocols find their applicability in disaster
area, military operation, etc., while non-DTNs routing
protocols is suitable for emergency response network which
gets further classified into beacon and beaconless protocols.
Some protocols uses bio processes as source of their inspiration
such ACO, BCO, swarm intelligence and genetic processes.
Some protocols based on ACO, BCO and genetic selection
processes were successfully simulated and implemented.
Routing in heterogeneous environment also poses some
challenges
like
handoff,
computational
complexity,
transmission time, mobility management and topology control
and thus we discussed some protocols which take these issues
into account.
Other aspects like implementation complexity, practical and
experimental popularity, density of vehicles, distinct
dimensions suitability and QoS can be delved to classify the
routing protocols. In future works we will try to cover these
aspects also for classifying VANETs routing protocols.
We also glanced at the various tools and technologies
available for the VANETs application‟s assessment. We
discussed mobility and network simulators. We also discussed
some integrated simulation environments like ITERIS and
Veins. In future works we will try to enlist some real world
experiments so far conducted in the realm of VANETs
applications.
5. References
W.H.O, Road traffic injuries, October-2015. [Online].
Available:
http://www.who.int/mediacentre/factsheets/fs358/en/
[2] H. Cai, L. Yan, D. Zhu, Using multi-regression to analyse
and predict road traffic safety level in China, in Proc. 3rd
Int. Conf. Transportation and safety, June 2015.
[3] N. F. G. C. Ferreira, J. A. G. Fonseca, Improving safety
message delivery through R.S.U‟s coordination in
vehicular networks, IEEE World Conference Factory
Commn. Systems, pp. 1-8, May 2015.
[4] Technical committee, CAR 2 CAR Communication
Consortium Manifesto, December 2007.
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[1]
[22]
[23]
C. Lochert, M. Mauve, H. Füßler, H. Hartenstein,
Geographic routing in city scenarios, ACM SIGMOBILE
Mobile Computing and Communications Review, volume
9, 2005, pp. 69-72
A. Bachir, A. Benslimane, A multicast protocol in ad hoc
networks inter-vehicle geocast, in: Proceedings 57th IEEE
Semiannual Vehicular Technology Conference, volume 4,
2003, pp. 2456-60.
C. Maihofer, R. Eberhardt, Geocast in vehicular
environments: Caching and transmission range control for
improved efficiency, in: Proceedings IEEE intelligent
vehicles symposium, 2004, 951-56.
H. P. Joshi, M. L. Sichitiu, M. Kihl, Distributed robust
geocast multicast routing for inter-vehicle communication,
in: Proceedings WEIRD workshop on WiMax, 2007, pp. 921.
M. Kihl, M. Sichitiu, T. Ekeroth, M. Rozenberg, Reliable
geographical multicast routing in vehicular ad-hoc
networks, in: Proceedings international conference WWIC,
2007, pp. 315-25.
C. Maihofer, T. Leinmuller, E. Schoch, Abiding geocast:
time-stable geocast for ad hoc networks, in: Proceedings.
ACM International workshop vehicular ad hoc networks,
2005, pp. 20-29.
C. Celes, C. T. De Oliveira, R. B. Braga, A. A. F.
Loureiro, R. M. C. Andrade, GeoSPIN: An approach for
geocast routing based on SPatial INformation in VANETs,
in: Proceedings. IEEE Vehicular Technology Conference,
2013, pp. 1–6.
H. Rahbar, K. Naik, A. Nayak, DTSG: Dynamic TimeStable Geocast Routing in vehicular ad hoc networks,
IEEE Computer Communication Symposium, 2010, pp.
198-203.
W. K. Wolterink, G. Heijenk, G. Karagiannis, Constrained
geocast to support cooperative adaptive cruise control
(CACC) merging,
in: Proceedings IEEE Vehicular
Networking Conference, 2010, pp. 41-48.
Y. S. Chen, Y. W. Lin, S. L . Lee, A mobicast routing
protocol for vehicular ad hoc networks, in: Proceedings
IEEE GLOBECOM , 2009, pp. 1-6.
S. Yu, G. Cho, A selective flooding method for
propagating emergency messages in vehicle safety
communications, in: Proceedings of International
Conference on Hybrid Information Techchnology, 2006,
pp. 556-61.
M-T. Sun, T-H. Lai, W-C. Feng, , H. Okada, K. Yamada,
GPS-based message broadcast for adaptive inter-vehicle
communications, in: Proceedings 52nd Vehicular
Technology Conference, 2000, pp. 2685-92.
A. M. Vegni, E. Natalizio, A. Stramacci, SRB: a selective
reliable broadcast protocol for safety applications in
VANET, in: Proceedings International Conference on
Selected Topics in Mobile & Wireless Networking, July
2012.
M. Durresi, L. Barolli, A. Durresi, Emergency broadcast
protocol for inter-vehicle communications, in: Proceedings
Intertational Conference Parallel and Distributed
Systems, 2005, pp. 402–06.
O. Tonguz, F. Bai, V. Sadekar, N. Wisitpongphan, P.
Mudalige, Broadcasting in VANET, in: Proceedings
Mobile Networking in Vehicular Environment, pp. 7–12,
2007.
W. Viriyasitavat, O. K. Tonguz, F. Bai, UV-CAST: An
urban vehicular broadcast protocol, in: Proceedings. IEEE
Vehicular Networking Conference, pp. 25–32, 2010.
J. Huang, Y. Huang and J. Wang, Vehicle density based
forwarding protocol for safety message broadcast in
VANET, The Scientific World Journal, pp. 1-9, 2014.
N. N. Nakorn, K. Rojviboonchai, DECA: Density-aware
reliable broadcasting in vehicular ad hoc networks, in:
Proceedings IEEE Computer Telecommunications and
Information Technology,pp. 598-602, 2010.
G. Korkmaz, Ü. Özgüner, F. Özgüner, E. Ekici, Urban
multi-hop
broadcast
protocol
for
inter-vehicle
communication systems,
in: Proceedings ACM
Intertational Workshop Vehicular Ad Hoc Networks, pp.
76–85, 2004.
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16398
DOI: 10.18535/ijecs/v5i5.12
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
G. Korkmaz, F. Özgüner, E. Ekici, An efficient fully adhoc multi-hop broadcast protocol for inter-vehicular
communication systems, in: Proceedings IEEE ICC, 2006,
pp. 423-28.
Y. H. Choi, F. Bai, P. Mudalige, R. Rajkumar, Adaptive
location division multiple access for reliable safety
message dissemination in VANETs, in: Proceedings 6th
Intertational
Conference
Symposium
Wireless
Communication System, pp. 565-69, 2009.
P. Akkhara, Y. Wakahara , Y. Sekiya, , Efficient alarm
messaging by multi-channel cut-through rebroadcasting
based on inter-vehicle communication, Intertational J.
Computer Science, May 2009.
W. Sun, F. Xia, J. Ma, T. Fu and Y. Sun, An optimal
ODAM-based broadcast algorithm for vehicular ad-hoc
networks, IEEE Transactions Internet and Information
System, December 2012, pp. 3257-74.
A. Benslimane, Optimized dissemination of alarm
messages in vehicular ad-hoc networks (VANET), in:
Proceedings IEEE International Conference High Speed
Networks Multimedia Communication, 2004, pp. 655-66.
H. Alshaer, E. Horlait, An optimized adaptive broadcast
scheme for inter-vehicle communication, in: Proceedings
IEEE VTC, pp. 2840-2844, 2005.
H. Jiang, , L. Chen, H. Guo, Reliable and efficient alarm
message routing in VANET, in: Proceedings ICDCSW,
2008, pp. 186-91.
T. W. Chen and M. Gerla, Global state routing : a new
routing scheme for ad-hoc wireless networks, in:
Proceedings
IEEE
International
Conference
Communication, 1998, pp. 171–175.
G. Pei, T. W. Chen, M. Gerla, T. W. Chen, Fisheye state
routing in mobile ad hoc networks, in Proceedings IEEE
International Conference Communication, 2000, pp. 7074.
V. D. Park, M. S. Corson, A highly adaptive distributed
routing algorithm for mobile wireless networks, in:
Proceedings IEEE INFOCOM, pp. 1405–13, 1997.
Z. J. Haas, "A new routing protocol for the reconfigurable
wireless networks," in: Proceedings IEEE Conference on
universal personal Communication, pp. 562–566, 1997.
N. Nikaein, N. Nikaein, C. Bonnet, HARP-hybrid ad hoc
routing protocol, in: Proceedings International Symposium
telecommunications, pp. 56–67, 2001.
V. Naumov, T. R. Gross, Connectivity-aware routing
(CAR) in vehicular ad-hoc networks, in Proceedings IEEE
INFOCOM, pp. 1919-27, 2007.
K. Prasanth, C. Chandrasekar, K. Jayasudha, K.
Duraiswamy, Minimizing end-to-end delay in vehicular ad
hoc network using edge node based greedy routing," in:
Proceedings
International
Conference
Advanced
computing, pp. 135-140, 2009.
H. Füßler, W. Effelsberg, H. Hartenstein, J. Widmer,
M.Mauve, ,Contention based forwarding for street
scenarios, in: Proceedings International WIT Workshops,
pp. 155-159, 2004.
Z. Mo, N. Pissinou, K. Makki, H. Zhu, MURU:a multi-hop
routing protocol for urban vehicular ad hoc networks, in:
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
Proceedings International Conference mobile and
ubiquitous systems - networking and services, pp. 1-8,
2006.
C. Lochert, H. Füßler, J. Tian, H. Hartenstein, D.
Hermann, M. Mauve A routing strategy for vehicular ad
hoc networks in city environments, in: Proceedings IEEE,
pp. 156–161, 2003.
R. Luy, X. Sheny, X. Linz , SPRING: A social-based
privacy preserving Packet Forwarding Protocol for
vehicular delay tolerant networks, in: Proceedings IEEE,
pp. 1-9, 2010.
C.Mascolo,
I.
Leontiadis,
Geopps:geographical
opportunistic routing for vehicular networks, in
Proceedings IEEE, pp. 1-6, 2007.
L. Boukhatem, G. Li, A delay-sensitive vehicular routing
protocol using ant colony optimization, in: Proceedings,
pp. 49–54, 2013.
J. C. Junior, O. Cherkaoui, S. L. O. B. Correia, Mobilityaware ant colony optimization routing for vehicular ad hoc
networks, in; Proceedings IEEE pp. 1125–30, 2011.
K-Y. Ho, C-H. Lin, C-H. Hsu, P-C. Kang, Implementation
of WAVE/DSRC devices for vehicular communications,
Computer Communication Control and Automation, 2010,
pp. 1-4.
David Padi, "Vehicular information and communications
technology (VICT) system," IEEE, 2009, pp. 1-5.
J.S. Huang, F. Lawal, O. Yang, "Sending Safety Video
over WiMAX in Vehicle Communications," in:
Proceedings Future Internet, pp. 535-567, 2013.
N. Jaber, N. C. Doyle and K. E. Tepe, " New combined
WiMAX/DSRC infrastructure design for efficient
vehicular networking," EURASIP Journal Wireless
Communication and Networking, 2012, pp. 1-26.
H. Haas, L. Yin, Y. Wang and C. Chen, "What is LiFi?,"
IEEE Journal Lightwave Technology, 2015, pp. 1-12.
K. Shafiee, V. C. M. Leung, A. Attar, “Optimal distributed
vertical handoff strategies in vehicular heterogeneous
networks,” IEEE Journal, pp. 534–44, 2011.
J. Kim, J-D. Cho, J. Jeong, J-Y. Choi, B-H. Song and H.
Lee, "Fuzzy logic based handoff scheme for heterogeneous
vehicular mobile networks, "IEEE Int. Conf. HPCS, 2014,
pp. 863-870.
N. Li and J. C. Hou, "in Proc. IEEE INFOCOM
Workshops, vol. 1, 2004, pp. 232-243.
DLR Institute of Transportation Systems, “Simulation of
Urban Mobility (SUMO),” [Online]. Available:
http://sumo.dlr.de/wiki/Main_Page
VanetMobiSim
project,
[Online].
Available:
http://vanet.eurecom.fr
CANU
project,
[Online].
Available:
http://canu.informatik.uni- stuttgart.de
STRAW as part of C3 project, “STreet RAndom Waypoint,”
2008 [Online]. Available:
http://www.aqualab.cs.northwestern.edu/projects/STRAW/
index.php
T. Sivakumar , IJECS Volume 05 Issue 5 May 2016 Page No.16383-16399
Page 16399