Cross Layer Approach for Efficient Dissemination of

Emergency Messages in VANETs

Mahadev A. Gawas, Palash Hurkat, Varun Goyal, Lucy J. Gudino
Department of Computer Science and Information Technology
Birla Institute of Technology and Science(BITS) Pilani K.K.Birla Goa Campus
Email id: {mahadev, lucy}@goa.bits-pilani.ac.in, { palashhurkat, varungoyalbits }@gmail.com

Abstract— This paper proposes a cross layer approach for governs the delivery ratio of the broadcast messages, especially
efficient dissemination of emergency messages in VANETs (CL- over densely populated networks [5]. The multimedia safety
DEM) by minimizing the message redundancy and maintaining low messages need high reliability and high priority to effectively
end-to-end communication delays. We propose a scheme to select a broadcast safety messages. Thus, the end-to-end delay for
one-hop neighbor relay as a potential forwarder for relaying the broadcasting messages should be minimized. Without any
broadcast messages to improve the transmission reliability in a control mechanism for broadcasting messages, flooding of
platoon of vehicles. The relay selection metrics compose of vehicular broadcast messages can occur between vehicles, because the
density, vehicular velocity, and the geographical location. The wireless channel is shared by all vehicles in the transmission
selected relay controls the broadcast messages with minimum
range. This can lead to poor network resource utilization,
overhead and with minimum bandwidth consumption. To provide the
service differentiation to different traffic classes, we adopt 802.11e
under-utilization of bandwidth, excessive re-transmission of
MAC. The cross layer is further extended to the transport layer to messages. To accomplish multi-hop data delivery and to
dynamically adapt the data transmission rate based on the physical prevent this message redundancy, VANET needs to select an
channel state. The extensive simulation analysis conducted reveals optimal next hop vehicle as the forwarder to disseminate safety
that the proposed cross-layer scheme effectively propagates the messages through a platoon which typically covers a few
critical broadcast messages with minimum latency. kilometers [6]. Various metrics can be used for selecting
suitable forwarder. For example, geographical information like
Keywords—Cross-layer; dissemination; emergency messages; the location of a vehicle through the GPS can be used to select
VANETs; a suitable forwarder. Most of the proposed multi-hop broadcast
protocols select the farthest vehicle in broadcast range as the
I. INTRODUCTION forwarder [7][8]. Although the end-to-end broadcast delay and
Cooperative communication in Vehicular Ad-hoc Networks the number of forwarding hops are decreased in that case, high
(VANET) has become the de-facto in enabling technology to contention delay is introduced in the case of high vehicle
improve its efficiency. Recently, the intelligent transportation density.
systems (ITSs) have been used in wireless communication to In this paper, we propose a cross-layer based distributed
enhance the road transportation systems. Most of ITS cooperative safety messages broadcast algorithm. The
applications utilize VANETs to provide the communications proposed scheme uses cross-layer coupling of the network,
between inter vehicular communication (IVC). VANETs are MAC, and physical layer to broadcast the messages through a
one of the promising wireless technologies that are largely cooperative relay in the network. The highlights of proposed
applied to enhance the driving safety and passenger comfort. CL-DEM are listed as follows:
VANETs can improve driving safety by dissemination of
safety and caution messages in the case of accidents. VANETs 1. The optimal one hop forwarder is selected based on
can also be used to avoid traffic congestion and provide multiple metrics like distance, position, and velocity
location services, thus increasing passenger comfort, reducing rather than selecting a random farthest vehicle as a
travel time, etc [2]. In recent years, although the progress in forwarder.
VANETs has evolved rapidly due to their diverse and useful
2. One hop forwarder vehicle is selected which is
applications, but several technical problems still need an
moving in the same direction as that of source by
attention. For example, guaranteeing the data delivery due to
computing its moving direction.
the high mobility of the moving vehicles, network topology
changes regularly, which leads to volatile links in inter-vehicle 3. To provide service differentiation to emergency
communication [3][4]. Different types of vehicles are present messages, 802.11e MAC is modified with an
on the road, each having contrasting processing power, the additional access category.
amount of memory present, the types of the antenna used,
which can lead to various different scenarios in a realistic 4. A variable data rate is chosen to disseminate
environment. Multi-hop messaging is the fundamental form of emergency messages rather than a base data rate.
communication in VANETs. Most of the existing IVC The next section will first review the related work. Section
techniques use a broadcast method to disseminate time critical III discusses the proposed protocol for next hop forwarder
safety information and selects the farthest vehicle in broadcast selection and a scheme for message dissemination. We provide
range as the forwarder to reduce the number of forwarding simulation results and conclusion in section IV and in section
hops. In the design of multi-hop broadcast messaging V, respectively.
technique, proper relay node selection is crucial, which highly

978-1-5090-4749-9/17/$31.00 ©2017 IEEE 206 ICUFN 2017

 II. RELATED WORKS inversely proportional to distance from the transmitter. Also,
Many research works have been studied in VANETs to the vehicle with lowest delay has the highest priority.
find the most effective way of transmitting emergency Another delay-based protocol ReC [15] uses geographical
messages through the network. However, still researchers are information by selecting the nearest vehicle to the centroid of
working to find a solution which is ideal for all scenarios. All neighbors that have not received the message as the
the methods proposed in the literature till today, come with its forwarding vehicle. It reduces the unnecessary transmission by
own drawbacks. Some of them include simple flooding which immediately retransmitting message received by selected
causes the broadcast storm problem in VANETs. The forwarder. However, its major drawback is to adapt quickly to
simultaneous broadcasting leads to frequent packet loss and high mobile environments.
collision due to contention in dense areas. In wireless
networks, achieving highly efficient multi-hop broadcast or III. PROPOSED PROTOCOL
coverage is a challenging task [4]. However, the broadcast
The proposed routing protocol disseminates a safety
problem can be solved by adjusting the delay or probability
message during the emergency situation from the source to the
which mitigates the channel contention brought by the
destination. During the routing process the optimal forwarder
flooding broadcast. Recently, some protocols have been
is chosen based on the vehicular density, vehicular velocity
proposed for emergency message delivery in IVC.
and geographical position of the vehicle. The proposed
In [9], Jagruti Sahoo et al., addresses the issue of the scheme uses cross-layer coupling of the transport, network,
emergency message dissemination in VANETs. The proposed MAC, and physical layer to broadcast the messages through a
protocol repetitively divides the area inside the transmission cooperative relay in the network.
using binary partition based approach to obtain the farthest
The considered network scenario with k types of vehicles
possible segment. The protocol adapts well to complex road
is in a multilane highway environment as shown in Fig.1.
structure and accomplishes directional broadcast for highway
Since the transmission range 𝑅𝑅, is much larger than the road
scenario. Another advantage of this protocol is the constant
width, the network scenario can be simplified as a one-
contention delay irrespective of the vehicle density.
dimensional VANET with road length of 𝐿𝐿. Each vehicle can
Samara et al., [10] have proposed Particle Swarm detect its location through GPS facility equipped with it. In
Optimization Contention Based Broadcast (PCBB) for fast and addition, the vehicle can also detect it’s one-hop neighbors
effective dissemination of emergency messages within a within its transmission range by exchanging a beacon
geographical area to distribute the emergency message and message.
achieve the safety system. It makes more accurate analysis and
A. Handshake Mechanism
increases the percentage of the emergency message reception
without affecting the channel collision. In the proposed protocol, to efficiently disseminate safety
In [11], Srinivetha R. et al., proposed a new adaptive message to optimal neighbor node, we perform a handshake
approach which uses the information about the urban procedure by exchanging REQB and RREB control packets
environment, thus increasing efficiency of warning message between adjacent one hop vehicles.
dissemination processes. It identifies vehicles in a dangerous Whenever the node has a critical safety message to
position and immediately sends warning messages to them. broadcast to vehicular traffic in the region, it broadcasts the
Javed Ma et. al., [12] proposed an efficient time-slotted REQB control packet and starts the timer TimerREQB. The
multi-hot broadcast protocol. This protocol selects a segment REQB packet composed of the source node address (S_info),
leader who is responsible for forwarding the warning message and its position(Sx and Sy), velocity (t_velocity ) which
on a particular road segment. It allocates separate time slots denotes current broadcast node’s moving velocity, Dmessage is
for warning messages to avoid interference. It ensures reliable the direction of the safety message, position ( Rx and Ry )
delivery through the signalling mechanism while maintaining which indicates the current broadcast node’s position, and the
high reception rate and low end-to-end delay for single-hop sequence number(Sseq).
safety messages. This protocol also reduces the number of
transmissions and ensures timely delivery of warning
messages.
In [13], a black-burst based ad hoc multi-hop broadcast
(AMB) protocol allows a neighboring node to send a channel
jamming signal with the time duration that is proportional to
its distance. However, it's major drawback is the long delay of
emergency messages caused due to largest jamming duration
used by the relay candidate.
An efficient 802.11-based protocol called urban multi-hop
broadcast (UMB) is proposed in [14]. UMB assigns
broadcasting delay to each node based on the distance between Fig.1. Demonstration of considered environment
the vehicle and transmitter. The broadcasting delay is

207
The neighbor nodes on receiving the REQB packet undergo
the qualifying test to be eligible for the relay node. The
neighbor nodes based on QoS metrics such as moving
direction, velocity, and vehicular neighbor density, determines
if it is suitable to act as a relay to propagate the broadcast
message efficiently. The neighbor node estimates its weight
Wrelay based on QoS metrics as given in (1).

( ) ( ) ( ) (1)

Where, and can be dynamically determined

Fig.2. Handshake sequence of control packets
according to the vehicle density and broadcast radius in
practical, ds-r is the distance between source and relay node, B. Packet Prioritization using 802.11E
ds-n is the distance between source and farthest node, ρ is the We propose a mechanism to update the priority of AC4
density of neighbor vehicles on a highway, ρmax is maximum with following modification to the EDCA function. In EDCA,
vehicular density, v is relative speed and V is maximum the contention window is used to calculate the number of time
speed. The neighbor node qualifies to be a relay if its Wrelay slots to backoff before gaining the channel access. The value
value is within the window of Wrelaymin and Wrelaymax. In order of the contention window (CW) is doubled every unsuccessful
to avoid the collision of REPB from multiple nodes qualified packet transmission due to collision, and the contention
to be a relay, each qualifying neighbor node waits for an window is reset to the minimum value for successful packet
addition λ time after sifs time. The value of λ is assumed transmission. By default, these values are statically set for
equivalent to difs. We divide the λ time slot into multiple τ slot each ACs. We have proposed a scheme in which increasing
of length σ. The σ and τ are computed as: CW limits on transmission failure or resetting due to success
is done in a gradual and non-uniform manner within the
(2) window range {CWmax[i], CWmin[i]}. For any unsuccessful
transmission of emergency messages due to collision from
⌊ ⌋ (3) AC4, our scheme performs the following measures. It checks
its current CW[i] value. If it is less than twice that of its
where, Pdelay is a channel propagation delay, and Tswitch is CWmin[i], its CW is increased at a faster rate by multiplying
transceiver switching time. We further divide the time slots τ by a factor of 1.5.
into ɸ to map Wrelay value, which is given as:
if (CW[i] < 2 CWmin)
CW[i]=min(CW[i]*1.5, 2*CWmin[i])
(4) Whenever there is a successful transmission of emergency
packets, we do not reset the CW[i] value to CWmin[i]. For
Each source node starts its ɸ counter from zero, and when AC4, contention window is linearly decremented if CW[i] is
reaches to its Wrelay value, sends REPB in that time slot. Thus, less than twice that of CWmin[i] otherwise it is decreased by a
a node with minimum Wrelay gets the channel access. The factor of 0.5.
REPB is composed of the sequence number, address, density,
and speed. The neighbor nodes eligible to be relay node, on if (CW[i] < 2 CWmin[i])
hearing the REPB packet from the neighbor node, it sets its CW[i]=max(CW[i]-1,CWmin[i])
stop its own back-off timer and update its NAV as per the Else
value included in the received REPB frame, and defer their CW[i]=max(0.5*CW[i],2*CWmin[i])
transmission accordingly. The source node on receiving REPB
within TimerREQB prepares for safety message broadcast. Fig.2 This assignment resulted in maximizing the probability of
demonstrates the handshake sequence of packets exchanged transmitting emergency messages for a given time. This
between the source and a relay. approach uses transport layer to regulate the volume of
To provide safety messages with higher priority than other emergency messages on the channel.
packets, we use the priority-based IEEE 802.11e MAC for The source node estimated the transmission rate DRi based on
service differentiation. In our proposed scheme CL-DEM, we the bit error rate (BER), estimated from the received signal of
modified the 802.11e (EDCA) to incorporate the fifth access REPB. The source node waits for SIFS time and broadcast the
category apart from existing four differential access categories message to the selected relay at a rate DRi. The relay replies
as shown in fig.3. The newly introduced fifth access category with the ACK on successful reception of the message.
(AC4) is reserved for emergency messages. Similarly, the selection of the next broadcast relay node is
executed. The neighbor nodes keep the record of overheard
safety messages to avoid redundancy and deal with traffic
congestion.

208
 Table I : Priority to Access category

User Priority Access category Designation

(UP) (AC)

1 0 Background

2 0 Background

0 1 Best effort

3 1 Best effort

4 2 Video

5 2 Video

6 3 Video
Fig.3. Modified 802.11e EDCA to accommodate emergency messages 7 4 Emergency Message

Algorithm: C. Proposed Cross Layer Architecture

Parameters: The traditional hierarchical TCP/IP layer architecture used in
{
networking is inflexible. The existing layered design is
: Distance between source and relay node
incapable of coping up with the dynamics of MANETs
: Distance between source and farthest node
supporting real time multimedia applications. In the proposed
ρ : density of neighbour vehicles
CL-DEM protocol, we propose a cross layer architecture
: Maximum vehicular density
framework, where information is passed across the layers to
υ : Relative velocity
V : Maximum velocity improve the network efficiency. The CL-DEM utilizes the
} geographical location and vehicle velocity information
Initialization: available at the physical layer for next hop forwarder selection
{ at the network layer. Based on the signal strength of received
= 0.4 REPB packet from relay node, the transport layer decides the
= 0.4 data rate for transmitting emergency messages. The MAC
= 0.2 layer is implemented with 802.11e EDCA functionality and
} modified to include a additional access category AC4, to
One-hop neighbor on receiving REQB message: accommodate emergency messages. Thus, giving higher
{ priority to emergency messages. The fig.4 explains the cross
Calculate from (1) layer approach proposed in CL-DEM algorithm.
If ( < < ):
IV. SIMULATION ENVIRONMENT
{
//node qualifies to be a relay The simulation platform used is NS-2.35 network
Calculate σ from (2) simulator [17]. Physical and MAC layer are simulated using
Calculate τ from (3) IEEE 802.11e inbuilt ns2 modules. The primary alterations
Wait(sifs) include 5.9 GHz as the carrier frequency and 10 MHz as the
width of the control channel band. Initially, there is a uniform
Wait for time slot
distribution of vehicles on a highway segment of
Send REPB
approximately 4 km length, having 4 lanes, the width of the
} each lane is 3.7 meters.
}
Source node on receiving REPB:
{
Wait(sifs)
Estimate the transmission rate 𝑅𝑅
Prepare for safety message broadcast
Insert the message in AC4 queue
}
One-hop neighbor on receiving the emergency message
{
Send ACK
Repeat the process for selection of next hop
forwarder
} Fig.4. Proposed cross layer architecture in CL-DEM

209
 All vehicles move with different velocity in the range of density is less, we receive slightly poor performance in terms
80-100 km/h, which is randomly assigned from the above- of the packet delivery ratio with the proposed method, as
mentioned range. Ten data flows are set up at the rate of 15 compared to FD and 802.11p. However, with an increase in
packets per second. The transmission range of the vehicles is the number of vehicles, CL-DEM outperforms the other
set to 300 meters along with a GPS device and a data popular protocols such as 802.11p and FD. The analysis also
processing unit. Also, there is a wireless transceiver which shows that when the number of vehicles exceeds 100, 802.11p
follows the IEEE802.11p standards. Packets of size 512 bytes provides the worst performance in the simulation scenario.
are emitted by CBR at an interval of 0.05 s. A platoon consists
of nodes varying in the range of 100 to 200, which increases
by 25 nodes per simulation scenario.
A. Performance evaluation
The proposed CL-DEM scheme is analyzed and compared
against two existing schemes 802.11p standard and
FD[18][19].
Fig.5. Shows the relay selection delays for the compared
protocols as a function of vehicle density variation. Relay
selection delay is defined as the time interval from which the
broadcast node attempts to deliver a REQB frame, to the time
it successfully receives a REPB frame. Service differentiation
is applied in the CL-DEM in which the emergency messages
are served with the highest priority. Alternatively, the FD ad
802.11p adopts a basic CSMA/CA mechanism, which doesn’t
consider the requirements of delay-sensitive traffic, and Fig.5. Relay selection delay v/s vehicle density
priorities of all the packets are same which access the channel.
Therefore, the access delay of FD and 802.11p is longer than
that of the CL-DEM. Additionally, the node sending the
longest channel jamming signal becomes the relaying node in
the FD, while a node waiting the shortest time to reply a
REQB frame becomes the relaying node in the CL-DEM.
The results of the broadcast count required to cover all the
vehicles are shown in Fig.6 The measurement of message
dissemination progress can be done by the broadcast count,
which considers both the success and failure broadcast. As a
result, two cases can be considered. For larger hop distances,
channel conflict probability will be high leading to a higher
number of failure broadcasts. However, for smaller hop
distances, coverage range is small leading to more relay hops.
802.11p randomly selects the forwarder vehicle and thus has
the highest broadcast count, but also requires more relay hops
Fig.6. Broadcast count v/s vehicle density
to transmit the message to last vehicles. If the objective is only
maximum coverage, then the broadcast count of FD performes
better than that of CL_DEM and 802.11p. The performance
of CL-DEM is nearly same as that of FD although CL-DEM
does not always choose the farthest vehicle as the forwarder.
Fig.7 shows the emergency access delay against the
vehicle density. Initially with lower vehicle density, all three
protocols incur minimum delay. As the density increases, the
result shows that CL-DEM outperforms FD and 802.11p.
This is because, firstly CL_DEM uses additional access
category by adjusting contention window parameters for
transmitting emergency messages with highest priority.
Secondly, the CL-DEM uses a higher data rate based on
channel quality to disseminate emergency messages. As a
result, the speed of dissemination of the message is reduced in
CL-DEM.
Fig.8 shows the packet delivery ratio for the compared Fig.7. Emergency message access delay v/s vehicle density
protocols. The figure analysis shows that, when the vehicle

210
 Transactions on Intelligent Transportation Systems, vol. 11, no. 1, pp. 90–
99, 2010.
[9] J. Sahoo, E. Wu, P. K Sahu, and M. Gerla, “Binary-Partition Assisted
MAC-Layer Broadcast for Emergency Message Dissemination in
VANETs”, IEEE Transactions on Intelligent Transportation Systems,
Vol. 12, No. 3, September 2011.
[10] S. Tareq Alhmiedat, “Intelligent Emergency Message Broadcasting in
VANET” using PSO of Computer Science and Information Technology
Journal (WCSIT) Vol. 4, No. 7, , 2014.
[11] R. Srinivetha., R. Gopi, “Inter Alert Message Dissemination Protocol for
VANET to Improve Road Safety” International Journal of Emerging
Technology and Advanced Engineering Volume 4, Issue 1, January 2014.
[12] M. A. Javed, N. DT, K. JY., “A multi-hop broadcast protocol design for
emergency warning notification in highway vanets”. Eurasip journal on
wireless communications and networking. 2014, Article 179.
[13] G. Korkmaz, E. Ekici, and F. Ozguner, “Black-burst-based multihop
broadcast protocols for vehicular networks,” IEEE Trans. Veh. Technol.,
vol. 56, no. 5, pp. 3159–3167, Sep. 2007.
[14] G. Korkmaz, F. ¨ Ozg¨uner, E. Ekici, and ¨ U. ¨ Ozg¨uner, “Urban multi-
Fig.8. PDR comparison v/s vehicle density hop broadcast protocol for inter-vehicle communication systems,” in
Proceedings of the 1st ACM International Workshop on Vehicular AdHoc
V. CONCLUSION Networks (VANET ’04), pp. 76–85,October 2004.
[15] J. Liu, Z. Yang, and I. Stojmenovic, “Receiver consensus: on-time
This paper proposed an effective multi-hop message warning delivery for vehicular ad-hoc networks,” in Proceedings of the
broadcasting scheme using a cross-layer approach in VANET 32nd IEEE International Conference on Distributed Computing Systems
called CL-DEM. The cross-layer approach blurs the gap (ICDCS ’12), pp. 386–395,Macau, China, June 2012.
[16] A. Amoroso , G. Marfia , M. Roccetti , Going realistic and optimal: a
between the layers to exploit the exchange of fine grain distributed multi-hop broadcast algorithm for vehicular safety, Comput.
information to enhance the communication in VANET. The Netw. 55 (10) (2011) 2504–2519 .
optimal selection of Relay with multi-QoS metric has a direct [17] Wireless and Mobility Extensions to the 13s-2 Network Simulator - CMU
impact on improving the packet delivery ratio over densely Monarch Project. http://monarch.cs.cmu.edu/cmu-ns.htm1.
[18] C. E. Palazzi, M. Roccetti, and S. Ferretti, “An Intervehicular
populated VANETs. The safety messages are given highest Communication Architecture for Safety and Entertainment,” Intelligent
priority compared to other control packets by adopting Transportation Systems, IEEE Transactions on, vol. 11, no. 1, pp. 90-99,
802.11e MAC. Also, messages are broadcast at the highest 2010.
possible rate based on channel state, minimizing the packet [19] W. Ben Jaballah, M. Conti, M. Mosbah, and C. E. Palazzi, “Fast and
Secure Multihop Broadcast Solutions for Intervehicular Communication,”
latency. Intelligent Transportation Systems, IEEE Transactions on, vol. 15, no. 1,
As a future continuation of this research, we would suggest pp. 433-450, 2014
investigating the performance of the CL-DEM scheme over a
variant of street topographies along with various obstacle
models.

REFERENCES
[1] M. A. Gawas, L. J. Gudino and K. R. Anupama, "Cross layered adaptive
cooperative routing mode in mobile ad hoc networks," 2016 22nd Asia-
Pacific Conference on Communications (APCC), Yogyakarta, 2016, pp.
462-469.
[2] F. Li, Y. Wang, Routing in vehicular ad hoc networks: a survey, IEEE
Veh. Technol. Mag. 2 (2) (2007) 12–22.
[3] W. Chen, K. Guha, J. Kwon, J. Lee, Y. Hsu, A survey and challenges in
routing and data dissemination in vehicular ad-hoc networks, in:
Proceedings of the IEEE International Conference on Vehicular
Electronics and Safety, Columbus, pp. 328–333, USA, 2008.
[4] Y.-C. Tseng, S.-Y. Ni, Y.-S. Chen, and J.-P. Sheu, “The broadcast storm
problem in a mobile ad hoc network,” Wireless Netw., vol. 8, no. 2/3, pp.
153–167, Mar. 2002.
[5] E. Hossain, G. Chow, V.C.M. Leung, R.D. McLeod, J. Misic, V.W.S.
Wong, O. Yang, Vehicular telematics over heterogeneous wireless
networks: a survey, Comput. Commun. 33 (2010) 775–793.
[6] B.T. Sharef , R.A. Alsaqour , M. Ismail , Vehicular communication ad
hoc routing protocols: a survey, J. Netw. Comput. Appl. 40 (2014) 363–
396 0.
[7] J. Huang , Y. Huang , J. Wang , Vehicle density based forwarding
protocol for safety message broadcast in VANET, Sci. World J. 2014
(2014) 9.
[8] C. E. Palazzi, M. Roccetti, and S. Ferretti, “An intervehicular
communication architecture for safety and entertainment,” IEEE

211