Wireless Personal Communications

https://doi.org/10.1007/s11277-019-06548-y

Internet of Vehicles Communication Technologies for Traffic

Management and Road Safety Applications

Arnav Thakur1 · Reza Malekian1,2,3 

© Springer Science+Business Media, LLC, part of Springer Nature 2019

Abstract
The primary objective of intelligent transport systems (ITS) is to improve road safety and
efficiency in the transport sector and are addressed in Internet of vehicles (IoV) based solu-
tions, peer to peer vehicle data sharing, inter vehicle and vehicle to infrastructure com-
munication channels. Effectiveness of an IoV solution is dependent on the robustness of
the wireless communication technology. Performance of ZigBee, Wi-Fi and DSRC com-
munication technologies for deployment in an IoV system is investigated by performing
simulations in an identified platform. It was found that ZigBee, Wi-Fi and DSRC can offer
successful exchange of data with failure rate of less than 1% for low frequency of commu-
nication events while Wi-Fi and DSRC can offer this performance at even higher frequen-
cies of exchange events. Findings of the research will be used to design and test road safety
and traffic management mechanisms for an IoV system.

Keywords  Internet of vehicles (IoV) · Vehicle to vehicle (V2V) communication · Vehicle

to infrastructure (V2I) communication · Wireless sensor network (WSN)

1 Introduction

Major challenges faced in the transport sector are focused around road safety and vehicular
congestion. The continually increasing trend of vehicular traffic has resulted in a drastic
reduction in the effectiveness of the transportation system as congestion results in increased
travel duration and hence an increase in fuel consumption that is adversely affecting air
quality [1]. The pressing need for a solution to effectively reduce vehicular congestion is
required to relief the imposed adverse environmental and economic effects associated with
the inefficiency in the transport sector [2].

* Reza Malekian
[email protected]
1
Department of Electrical, Electronic and Computer engineering, University of Pretoria,
Pretoria 0002, South Africa
2
Department of Computer Science and Media Technology, Malmö University, 20506 Malmö,
Sweden
3
Internet of Things and People Research Center, Malmö University, 20506 Malmö, Sweden

13
Vol.:(0123456789)
 A. Thakur, R. Malekian

The use of location based alerts, which provide motorists with traffic related informa-
tion alerts via various mobile and online platforms such as mapping services are used by
traffic management systems in an attempt to reduce vehicular congestion [3, 4].
These systems are limited by the effective coverage area as well as the costs of imple-
menting the system using exorbitant cellular networks, particularly in remote areas. Thus,
a dedicated traffic detection and management system that enhances the utilization and
throughput of road infrastructure is required for improving the traffic flow by providing the
motorists with real time location based alerts relating to recommended alternative routes as
well as driving speeds [1, 3].
Vehicular accidents are a common phenomenon that is responsible for human injury
and damage to property. The fact that a human fatality resulting from a vehicle accident
occurred once every 15 min in the US itself highlights the adverse impact vehicle accidents
have on human life [5]. It was found that 60% of all vehicular collisions could be prevented
if contextual preventive measures were provided to the vehicle operator as little as 1.5  s
before a possible event [5].
The use of vehicle mounted sensors are implemented by current road safety systems for
assisting motorists by providing blind spot assistance, maintaining safe following distances
and lane assistance but are incapable of detecting non line of sight factors of collisions
[6]. The results from the data analyzed by the California performance measurement system
show that with the use of current technologies, safe inter-vehicle distance can be optimized
to such an extent that vehicles only utilize 5% of the highway surface [7]. Thus, the need
for a road safety system implementation is of vital importance for aiding in the reduction of
vehicular accidents by providing preemptive contextual alerts and advisories through inter
vehicle communication channels and peer to peer data sharing mechanisms.
The Internet of vehicles (IoV) is an Internet of Things (IoT) based solution targeted at
meeting the road safety and traffic management goals of intelligent transportation systems
(ITS) for vehicular clients which has gained headway as a result of the increased reliability
of wireless communication and sensor technologies [6, 8]. The IoV solution consists of a
wireless sensor network (WSN) containing a central server which processes the received
data in real-time to generate contextual measures, vehicular client nodes and roadside units
(RSU) which function as waypoints for collecting the required sensory data and dissemina-
tion of safety and traffic related alerts [6].
In this paper, IoV based system for road safety and traffic management applications is
studied and the communication channel is investigated. A comparative study of simulation
frameworks for designing and testing of road safety and traffic management related appli-
cations for services used in IoV solutions are studied. Communication channels are studied
by simulating multiple wireless communication technologies. Requirements such as trans-
mission rates and message size for the communication channels is identified. Performance
of the wireless communication technologies is used to determine the most suitable wireless
communication technologies for the communication channels in an IoV system.
The remainder of the paper is organized as follow: In Sect. 2, IoV architecture is stud-
ied and the performance requirement of the communication channel is identified for road
safety and traffic management applications. Various network simulation platforms for
vehicular WSNs are studied in Sect.  3. Appropriate simulation platform for testing IoV
systems is identified in Sect. 4. Communication channels in the IoV based system are stud-
ied by performing simulations in Sect. 5. Multiple parameters are used to study the com-
munication channels and results obtained from the simulations undertaken in this study is
summarized in Sect. 6. The obtained results are discussed in Sect. 7 and a conclusion is
drawn in Sect. 8.

13
Internet of Vehicles Communication Technologies for Traffic…

2 IoV System

An Internet of Vehicles (IoV) systems consists of a wireless sensor network containing

vehicular client nodes, stationary waypoints and a centralized server [9]. Stationary way-
points function as roadside units (RSU) which are used for collection of sensory data from
the vehicular nodes and dissemination of alerts to client motorists generated on the server.
Apparatus to collect sensory data from vehicular client nodes along with the communica-
tion system is contained in the vehicle onboard unit (OBU) [3, 10]. The architecture of an
IoV WSN is illustrated in Fig. 1.
IoV based WSN utilize inter-vehicle communication channel referred as vehicle to
vehicle communication (V2  V) and communication between RSUs and vehicle OBUs
referred as vehicle to infrastructure communication (V2I) [3–5, 11]. Road safety related
warnings and alerts relevant to peer motorists are exchanged using V2 V communication
[8]. Whereas V2I communication is used for collection of sensory data and dissemination
of location based contextual alerts to client motorists [12]. Wireless communication tech-
nologies for the channels include cellular networks, DSRC/WAVE, Wi-Fi and ZigBee [10,
13–16].
Traffic management systems in IoV solutions use traffic status information to identify
vehicular congestion by using techniques such as adaptive speed limits and optimized
signal control [17, 18]. The techniques are aimed at improving the efficiency and utiliza-
tion of the infrastructure by enhancing traffic flow and road occupancy rates. Sensory data

Fig. 1  Architecture of WSN used in IoV systems consisting of vehicle OBU, RSU and central control unit
that use V2 V, V2I communication channels. (From [9])

13
 A. Thakur, R. Malekian

collected from vehicular nodes include frequency of braking events, average acceleration,
speed and position [19, 20]. Traffic status is identified by using sensory data collected from
vehicular nodes which is processed by the central control unit by using statistical methods,
vehicle beacon frequency and clustering algorithms [2, 21, 22].
Sensors such as accelerometers and gyroscopes are used to detect events which can
compromise road safety. Peer to peer data sharing to avoid collisions that can be caused
by negligent driving, improper driving practices and lack of maintaining of safe driving
distance [6, 13]. Road safety related guidance is also provided via location based alerts [8,
13].
A variety of message packets containing sensory data and processed responses are
exchanged in the IoV system through the V2 V and V2I communication channels for the
described road safety and traffic management mechanisms. A basic set of applications for
intelligent transportation systems has been defined by the European Telecommunications
Standards Institute (ETSI).
Information exchanged in the IoV is classified by ETSI into categories and the mini-
mum frequency of the messages based on the periodic exchange of information along with
the communication channel used is summarized in Table 1. The maximum communication
latency for the exchange of message packets is 100 ms [23]. This is used to evaluate the
performance of the wireless communication technologies for usage in the IoV WSN by
conducting a simulations based study in the identified platform.

3 Overview of Simulation Platforms

Network simulation platforms enable evaluation of multiple topologies and schemes effec-
tively, as a test bed based behavior analysis approach can be a time consuming exercise
with high input costs [24, 25]. Network simulation platforms can include hardware and
software components which can predict the behavior of a network without the presence of
an actual network [25]. Testing of the behavior and mechanism of the proposed network
based applications can be done before deploying it in a network. Thus performance of the
design can be verified [26]. Network based simulations are similar to a test bed as it ena-
bles direct experiments to be conducted for studying the impact of alternations of system

Table 1  Summary of minimum recommended frequencies for message categories from the ETSI basic set
of applications for ITS [23]
Category Application Minimum fre- Commu-
quency (Hz) nication
channel

Road safety Emergency braking 10 V2 V

Road safety Abnormal driving conditions 1 V2I
Road safety Collision warning 10 V2 V, V2I
Road safety Cooperative awareness 2 V2 V
Road safety Lane change assistance and overtaking warning 10 V2 V
Traffic management Traffic hazard warning 10 V2I
Traffic management Contextual driving guidance and alerts 1 V2I
Traffic management Intersection management 1 V2I

13
Internet of Vehicles Communication Technologies for Traffic…

parameters on performance [25]. Graphical and trace file based results can be obtained
from network simulators. Every event of the simulation can be recorded and parameters
such as throughput, packet delivery time, round trip time and communication latencies can
be collected for analysis [25, 27].
Platforms such as NS-2, NS-3, OPNET, OMNET ++, J-Sim, QualNet can be used for
simulation of network based applications [24–26, 28]. Network simulators are studied in
the following to identify simulation frameworks for testing IoV related algorithms and
schemes composed of vehicular networks for road safety and traffic management related
applications.

3.1 Network Simulator (NS‑2)

NS-2 is a open source C ++ based discrete event simulator [27]. Networks containing wired
and wireless channels of arbitrarily defined topologies using different protocols can be sim-
ulated [24]. NS-2 uses C ++ source code along with OTcl script. The simulation interface
is provided by OTcl, a Tcl based objected scripting language [25]. A queuing system is
used to process the scheduled network events. The simulations in NS-2 are virtual [24].

3.2 Ns‑3

Network simulator (NS-3) is an independent discrete event simulator that does not sup-
port Ns-2 APIs [1]. The programming languages used in the simulation framework is a
combination of C ++ and python [1, 3]. Parallel simulations along with emulations can be
performed in the realistic environment provided by the platform [1]. Additional advantages
of the platform include detailed models for protocols such as IEEE 802.11 [4]. The support
of only IPv4 addressing is a disadvantage of this platform [3].

3.3 Opnet

OPNET is a C ++ based discrete event simulation platform for high level network level
simulations [25, 29]. Explicit models are provided by the platform to simulate messages
and protocols [25]. Distributed systems and communication networks can be modeled in
the development environment [29]. Network behavior of the model can be studied by using
the flow analysis feature of the platform [25].

3.4 OMNeT  ++

OMNeT ++ is an open source discrete event simulation platform with a flexible architec-
ture which enables it to be used in studying of queuing systems, networks, distributed com-
puter hardware and evaluating performance of complex software systems [24, 27–29]. It
contains an execution environment, simulation kernel and an IDE [30]. Behavior of the
model is described using C ++ based source files. Composition of the model and relation-
ship between the elements of the model are described using the network description (NED)
language [25]. The platform allows reusable models as it follows the structured compo-
nent oriented approach [27]. The extensive graphical interface enables detailed analysis of
events and parameters of the model [24, 27].

13
 A. Thakur, R. Malekian

3.5 J‑Sim

J-Sim is a Java component based simulation platform [25, 26]. The platform uses the
autonomous component architecture which enables independent component based design
and implementation [25]. The platform is versatile as it supports integration with multiple
scripting languages such as Python, TCL and Perl and has a reusable simulation environ-
ment [29]. Disadvantages of the platform include limited support to MAC protocols for
wireless networks and high runtime overhead which make the simulations inefficient [27].

3.6 QualNet

QualNet is a simulation platform that uses Parsec C ++ and is developed especially for
large scale networks [25]. It can be uses to predict performance of wireless and wired
networks and networking devices [29]. Real-time speed for hardware modeling and soft-
ware emulation are supported by the platform [26]. The platform is a commercial product
derived from GloMoSim and so is not ideal for research [26].

4 Identification of Simulation Platform

With the constraints of mobility and motion in vehicular networks, the criteria for selecting
a simulation platform include scale and applicability along with support for multiple net-
work protocols and architectures [30]. The simulation platform has to be scalable to simu-
late the dimensions of the network area and number of vehicular nodes in the scenario.
The models of the simulation platform are required to be applicable and should be able to
represent factors such as the communication channel, MAC protocol and network latencies
of the scenarios close to field conditions [30].
Since J-SIM has limited support of wireless MAC protocols and inefficient simulations
due to high run-time overhead, it was not considered for simulating networks applications
in IoV systems as the requirements are not met. High memory consumption of models
in OPNET and inadequate models for vehicular network scenario make it inadequate for
simulating IoV systems.
A wide range of models ranging from IEEE 802.11, 3GPP LTE to MANET routing pro-
tocols are supported by NS-2 and NS-3 [30]. NS-3 simulates transmission data at packet
level and a module can be used for physical layer implementations. The module consumes
a significant amount of computing resources and is restricted to small scale wireless com-
munication scenarios [30].
OMNeT ++ modules support simulation of peer to peer networking and wireless com-
munication technologies such as LTE, LTE-A, IEEE 802.11, IEEE 802.11p (DSRC) and
IEEE 1609.4 (WAVE) [30]. The models are configured to simulate path loss due to factors
such as interference, shadowing from buildings and vehicular motion [30].
Performance of OMNeT ++, NS-2 and NS-2 are studied in [24]. The parameters
evaluated are memory usage, CPU utilization and computation time for a simulation
scenario with 400–2000 nodes and packet size of 512 kb exchanged between the nodes.
It was observed in the study that the highest amount of memory consumed was by NS-2
and NS-3 consumes the least. CPU utilization for NS-2 and NS-3 was found to be 50%
for the scenario and for OMNeT it was 35%. The computation time for NS-2 increases

13
Internet of Vehicles Communication Technologies for Traffic…

Table 2  Comparison of network simulation platforms [24, 25, 27, 29, 30]

NS-2 NS-3 OMNeT

Programming language C ++, OTcl C ++, Python C ++, NED

Scalable No Yes Yes
Simulation time Poor Fast Moderate
Wireless communication tech- IEEE 802.11 IEEE 802.11, 3GPP LTE IEEE 802.11, LTE,
nologies model supported LTE-A, DSRC/
WAVE

Fig. 2  Hierarchy of Veins framework for the OMNeT ++ platform for simulating road safety and traffic
management related IoV applications (From [30])

with the highest gradient for increasing number of nodes when compared to OMNeT
and NS-3. Thus NS-2 is not scalable and cannot be used to simulate scenarios with large
number of nodes. NS-3 is the fastest simulator. Whereas OMNeT can be used for large
networks with most realistic models representing communication channels and informa-
tion flow most accurately. The above simulation platforms are summarized in Table 2.
From the studied simulators, OMNeT ++ is identified as an ideal simulator for simu-
lating IoV networks. Advantages include the ability to model large networks and sup-
port for wireless communication such as IEEE 802.11, DSRC and WAVE used in IoV
systems. Veins (Vehicles in Network Simulator) is an inter-vehicle communications
framework for OMNeT [30]. This enables the OMNeT platform to provide a realistic
simulation environment which can model vehicle movement and flow for testing IoV
system related applications of road safety and traffic management. The Veins simulation
framework for the OMNeT simulation platform is described in Fig. 2.

13
 A. Thakur, R. Malekian

5 IoV Simulations

Wireless communication technologies that can be used in V2 V and V2I communication
channels are simulated in the OMNeT ++platform for communicating nodes in motion.
Multiple parameters are used to evaluate the performance with the ETSI requirements of
minimum message frequency and maximum communication latency. Successful recep-
tion of packets sent for various scenarios are recorded to identify the most appropriate
technology for V2 V and V2I communication. Transmission of packets over the wireless
communication channel used in this study is shown in Fig. 3.
Wireless communication technologies that offer a range of at least 100 m are studied.
This includes ZigBee, Wi-Fi, Dedicated Short Range communications (DSRC). Model
dimensions, modeling of nodes in motion, simulation time, packet size, traffic model
and wireless communication channel are aspects pertaining to the simulation and are
described in the following.

5.1 Model Dimensions

The wirelessly communicating network nodes in the model are placed in a test environ-
ment of size 1100 × 500 m for the channel using Wi-Fi and DSRC. Since ZigBee offers
a shorter coverage area of just 100  m, the model uses a smaller test environment of
dimensions 300 × 300 m. In the initial condition, the communicating sender and receiv-
ing nodes are separated by a distance of 10 m.

Fig. 3  Model used for wireless

communication between hosts
simulated in the OMNeT ++ plat-
form

13
Internet of Vehicles Communication Technologies for Traffic…

5.2 Modeling Motion of Communicating Nodes

Modeling of motion of communicating nodes is achieved by using the linear mobility sub-
module of the INET framework of the OMNeT ++ platform. The sender node moves at the
specified speed parallel to the receiving node for the duration of the simulation. Determin-
istic movement in a straight line is used as vehicles travel in a Manhattan grid road topol-
ogy [31].

5.3 Simulation Time

The model is simulated for a fixed duration. The simulation time is equivalent to the time
taken by the node travelling at a fixed speed to cover the range of the wireless communica-
tion technology of the channel. The simulation time for DSRC and Wi-Fi with a range of
1 km varies from 250 to 30 s for speed of the host nodes ranging from 15 to 120 km/hr.
Simulation time for ZigBee with a range of 100 m varies from 25 to 3 s for host nodes in
motion with a speed in the range of 15 to 120 km/hr.

5.4 Packet Generation

In an IoV system packets traverse distances of less than 1 km and thus there is negligible
propagation delay. Transmission delay is required to be less than 100 ms to meet the ETSI
requirements for road safety and traffic management related messages. Transmission delay
is the ratio of the length of the packet being transmitted and the data rate of the channel
and is described in Eq. 1 [32]. Maximum packet size that can be exchanged for the ZigBee,
Wi-Fi and DSRC to meet the communication delay requirement are calculated using Eq. 1
and is summarized in Table 3 [32].
Length of packet(L)
Transmision delay = (1)
Data rate(R)

5.5 Traffic Model

Timing of exchange of data packets in an IoV system follow the memory-less exponen-
tial distribution from a Poisson process, where the mean is the frequency of the messages
being exchanged [5]. In the simulation, messages are generated randomly where the time
of generation of packets follows an exponential distribution with a mean frequency rang-
ing from 1 Hz to 1 kHz. Packet size used in an IoV system can have a size of 300 Bytes
[31]. Using the results obtained in Table 3, a packet size of 300 Bytes is used for Wi-Fi and
DSRC simulations. A packet size of 300 Bytes is used for ZigBee simulations for message

Table 3  Maximum packet size Communication Data rate Maximum packet size

for message frequency of 1 Hz technology for message frequency
to have a maximum transmission of 1 Hz
delay of 100 ms
ZigBee 250 Kb/s 3125 B
Wi-Fi 56 Mb/s 700 KB
DSRC 27 Mb/s 337.5 KB

13
 A. Thakur, R. Malekian

frequencies of 1–10 Hz, 31 Bytes for 100 Hz and 3 Bytes for 1000 Hz as it has lower data
rates. Counters are used to keep track of the number of packets sent and received success-
fully. The percentage of packets received successfully for the scenario being simulated in
the test is used for comparing the wireless communication technologies.

5.6 Communication Channel

The wireless hosts in the model are of the standard NED host type. The host type contains
generic templates for UDP and TCP/IP routing. In the simulation, data packets are sent to
the receiving host wirelessly over a UDP stream. UDP stream is used in the simulation as
transmission time is critical in ad hoc networks used in vehicular WSN [10]. Assigning of
the IP addresses and setting up of static routing is done by using the IPv4Configuration
and port 4000 is used by the 2 host nodes in the model. IP address ranges are specified
in a XML string and the route with the least packet error is used for transmission of the
packets. Queue size of the wireless interfaces is fixed at 1 in this study to determine maxi-
mum number of packets that could be dropped due to congestion related to communicat-
ing nodes in motion. This was also done to have no queuing delay in the communication
channel. The software defined radio for transmitting and receiving packets uses amplitude
and phase shift keying with the BPSK modulation scheme The wireless configuration of
the software defined radio used in this simulation for DSRC is based on the US ASTM
standard [10]. Wireless configuration for the software defined radio used to model com-
munication at the physical layer for ZigBee, Wi-Fi and DSRC used in this simulation is
summarized in Table 4.

6 Results

The experiments undertaken in this study were to observe percentage of packets success-
fully received for communicating nodes in motion for the cases with varying speed of the
sender node with fixed packet size, varying packet size and varying frequency of exchang-
ing of messages. This was done to determine the performance of the wireless communi-
cation technologies in dynamic situations for determining the most robust option for the
V2 V and V2I communication channels used in an IoV system for exchanging time sensi-
tive safety related information. Results obtained from the experiments in this study is as
follows.

Table 4  Software defined ZigBee Wi-Fi DSRC

radio configuration used for
the wireless communication
Standard IEEE 802.154 IEEE 802.11 IEEE 802.11p
technologies in the simulation
to model the physical level [3, 7, Carrier frequency 2.4 GHz 2.4 GHz 5.9 GHz
10, 15] Channel bandwidth 2 MHz 20 MHz 10 MHz
Data rate 250 kb/s 56 Mb/s 27 Mb/s
Range 100 m 1 km 1 km
Transmission power 80 mW 100 mW 760 mW
Simulation environ- 300 × 300 m 1100 × 500 m 1100 × 500 m
ment dimensions
Simulation time 3 to 25 s 30 to 250 s 30 to 250 s

13
Internet of Vehicles Communication Technologies for Traffic…

6.1 Varying Speed of Sender Node

In this experiment, messages are sent to the receiver by the moving sender node. Messages
are generated at random with a mean frequency of 10 Hz for ZigBee and1 kHz for Wi-Fi
and DSRC. Packet size is kept constant in this experiment. Packet size used for DSRC and
Wi-Fi is 300 Bytes and 300 Bytes for ZigBee to have a maximum transmission delay of
100 ms. The experiment was conducted for a stationary receiver and a receiver at rest. The
speed of the sender is varied form 15 to 120 km/hr. Thus, the relative speed for the receiver
with respect to the source is a maximum of 120 km/hr. Results obtained from this experi-
ment is shown graphically in Fig. 4.
From the results obtained in this experiment, it can be observed that some packets are
dropped during the transmission. The minimum percentage of packets dropped is 2.40%
while using the Wi-Fi configuration for a relative speed of 15 km/hr. The maximum per-
centage of packets dropped is 14% for communication using DSRC with relative speed of
120 km/hr.
It can be observed that with increasing relative speed, the percentage of packets being
dropped increases. For Wi-Fi, the percentage of packets successfully transmitted ranges
from 97.5 to 91% for relative speeds from 15 to 120 km/hr and the average packet loss is
5.6%.
DSRC has the highest packet loss with an average of 12%. The minimum packet loss is
8.7% for relative speed of 15 km/hr and a maximum packet loss of 14% for a relative speed
of 120 km/hr.
Packets dropped while using ZigBee varies from 2.6 to 6.4% with an average loss of
4.5%. A fluctuating trend can be observed with ZigBee as the amount of packets sent
is extremely low when compared to DSRC and Wi-Fi. Frequency of messages being
exchanged with ZigBee is 1% of the frequency of messages for DSRC and Wi-Fi. Lower

Fig. 4  Percentage of packets received successfully for relative speeds between 15 and 120 km/hr

13
 A. Thakur, R. Malekian

frequency of information being exchanged with ZigBee due to lower bandwidth leads to
lower congestion when compared to DSRC and Wi-Fi. Thus, ZigBee has the least amount
of packets lost in this experimental scenario.
Packet loss in this experiment is not only because of congestion caused due to motion
but can also be attributed to the Doppler Effect phenomenon experienced. The effect occurs
when the receiver of the wave and the source of the wave are in motion with respect to each
other, which causes alteration of the received frequency. Frequency is gained for the source
moving closer to the receiver and is lost for the source moving away from the receiver [33].
The alterations in received frequency due to the Doppler Effect is given in Eq. 2.
Δv
( )
Frec = fc 1 ± , (2)
c
where Frec is the received frequency, fc is the carrier frequency. Δv is the difference in
speed and c is the speed of light.
Majority of the sent packets are still received successfully as the loss in frequency for
speeds as high as 120 km/hr is less then channel bandwidth for the three wireless commu-
nication technologies.

6.2 Varying Packet Size for Moving Node

In this experiment, percentage of packets received successfully for the sender node in
motion is studied. Messages with packet sizes ranging from 500 Bytes to 3 Kbytes are sent
to the receiver over the UDP stream. Packets to be transmitted are generated at random
intervals with a mean frequency of 1 kHz. The sender node’s speed in this experiment is
kept constant at 60 km/hr. The wireless communication technologies studied in this experi-
ment are DSRC and Wi-Fi. ZigBee was not considered for this experiment as it offers low
data rates and it can only transfer a fraction of messages at high frequency compared to
DSRC and Wi-Fi in the same time period. Thus, the comparison of ZigBee with DSRC and
Wi-Fi will not be equitable for the test scenario of this experiment. Results obtained from
this experiment is shown graphically in Fig. 5.

Fig. 5  Percentage of packets received successfully for message size ranging from 500 Bytes to 300 Kbytes
with sender node in motion with a constant speed of 60 km/hr

13
Internet of Vehicles Communication Technologies for Traffic…

With increasing size of the packets, the percentage of packets received successfully
reduces. The minimum packet loss is 5.37% for Wi-Fi with packet size of 500 Bytes. A
maximum packet loss of 38%occurs with DSRC for a packet size of 3 Kbytes.
Wi-Fi has an average packet loss of 16.87%. The minimum packet loss is 5.37% for
a packet size of 500 Bytes and the maximum packet loss is 26.18% for a packet size of
3 Kbytes. DSRC has a high average packet loss of 26.7%. The minimum packet loss is
11.87% for a packet size of 500 Bytes and the maximum packet loss is 38.34% for a packet
size of 3 Kbytes.
The network simulated in this experiment is a packet switched network which suffer
packet loss due to link congestion [32]. Packet loss can be minimized by increasing the
queue lengths of the networking interfaces and increasing transfer rates which also increase
throughput of the link [32].
Lower packet loss experienced for Wi-Fi in this experiment is because it offers twice the
data rate compared to DSRC. Higher data rates ensure that transmission delay with Wi-Fi
is lower than DSRC and thus congestion is minimized, as well as increasing the throughput
of the link.

6.3 Varying Message Frequency for Moving Node

In this experiment, percentage of packets received successfully for varying frequency

of messages being sent is studied for a sender node with a constant speed of 60  km/hr.
ZigBee, Wi-Fi and DSRC are the wireless communication technologies compared in this
experiment. Messages are generated at random intervals with a mean time ranging from
1 Hz to 1 kHz. Message size exchanged for the varying frequencies is chosen to meet the
maximum ETSI transmission delay requirement of 100 ms. Thus, the message size used for
ZigBee and Wi-Fi is 300 Bytes. The message size used for ZigBee is 300 Bytes for a fre-
quency of 1 and 10 Hz, 30 Bytes for a frequency of 100 Hz and 3 Bytes for a frequency of
1 kHz. This is because of the lower data rate offered by ZigBee. The results obtained from
this experiment for ZigBee, Wi-Fi and DSRC is compared graphically in Fig. 6.
As the frequency of the messages increases, the percentage of packets delivered
decreases. The average percentage of packets dropped by ZigBee in this experiment is
17.29%. For Wi-Fi the average percentage of packets dropped is 1.5% and for DSRC it is
3.3%.
The rate of successful delivery of packets decreases with the fastest rate for ZigBee
and the slowest is for Wi-Fi. For a frequency of 1 Hz, no packets are dropped for all the
three wireless communication technologies. For a frequency of 10 Hz, percentage of pack-
ets dropped is less than 1% for all the three wireless communication technologies. For a
frequency of 100  Hz, a sharp decline in performance of ZigBee can be noted where the
percentage of packets dropped is 14.8%. For the same frequency Wi-Fi and DSRC loose
less than 1% of the packets being sent. Wi-Fi looses only 0.26% packets and DSRC looses
0.74%. For a frequency of 1 kHz, ZigBee shows poor performance as 50% of the packets
are not delivered. Wi-Fi has the highest success rate for a frequency of 1 kHz as only 5.4%
packets are dropped whereas DSRC looses 11.9% of the transmitted packets.
A high frequency of messages being sent causes link congestion which results to
increase in the number of packets not being delivered successfully. From the results of this
experiment, it is found that ZigBee is not suitable for applications requiring high frequency
of messages being sent as Quality of Service (QoS) is unacceptable. DSRC and Wi-Fi on

13
 A. Thakur, R. Malekian

Fig. 6  Percentage of packets received successfully for message frequency of 1  Hz to 1  kHz with sender
node in motion with a constant speed of 60 km/hr

the other hand can be used for applications where large amounts of packets are exchanged
in a short duration.

7 Discussion

Internet of Vehicle mechanisms exchange messages that contain sensory data and pro-
cessed responses for enhancing road safety and efficient traffic management. The fre-
quency of messages exchanged in the system and the maximum packet size depend on the
mechanism of the application. Effective IoV solutions in ITS for road safety and traffic
management are dependent on the reliability and performance of the V2 V and V2I com-
munication channel [3, 12, 34]. Thus, choosing the most appropriate wireless communica-
tion technology is critical for development and testing of a practically feasible IoV system
designed for road safety and traffic management applications.
Since test bed based approaches have high input costs, a simulation platform is used as
it can predict the behavior of the network in the communication channel without its physi-
cal presence [25]. OMNeT ++ was the simulator used in this study because of its ability to
represent realistic test environments and support for Veins that can be used for simulating
inter vehicle communication systems and traffic management applications [25].
From the simulations it was observed that ZigBee, Wi-Fi and DSRC are capable of pro-
viding communication links between moving transceivers for low frequency of messages
with a success rate of 100%. Success rate of exchange of data with Wi-Fi for communi-
cating nodes in motion is the best as it has the highest data rates and channel bandwidth
allocated in comparison to DSRC and ZigBee and is therefore the least affected due to the
alternations in received frequency due to the Doppler Effect.
ZigBee cannot be used for exchanging packets of size greater than 300 Bytes at
frequencies higher than 10  Hz as transmission delay will exceed the 100  ms ETSI

13
Internet of Vehicles Communication Technologies for Traffic…

communication latency requirement due to low data rates offered and QoS requirements
are not met as packet loss is more than 1%. ZigBee however meets the ETSI require-
ments for the road safety and traffic management applications summarized in Table 1 for
exchange of packets at lower frequency.
DSRC and Wi-Fi are capable of transferring messages with packet size of 300 Bytes
at a frequency of 100  Hz with a failure rate of less than 1% and a maximum trans-
mission delay of less than 100 ms. For higher frequency of exchanging of information,
Wi-Fi is more suitable. But the failure rate is greater than 1% and can lead to compro-
mising quality of service. Wi-Fi and DSRC offer low percentage of packets dropped
for high frequency of exchange events and are capable of transferring large amounts
of data. Thus, DSRC and Wi-Fi can be used for V2I communication channels [31].
ZigBee cannot reliably transfer large amounts of data and thus is unsuitable for V2I
communication.
In a simulation based study conducted in [5], it was found that ZigBee and Wi-Fi are
capable of successfully exchanging message packets of size 300 Bytes between moving
transceivers at speeds as high as 120 km/hr.
It was found in [35] that DSRC can successfully transfer messages with packet size
of 250 Bytes with a communication latency of less than 100  ms between transceivers
moving with a relative speed of 120  km/hr. On the other hand, for the same scenario,
communication latency exceeded 1400  ms and thus, LTE based cellular networks do
not meet the ETSI requirements and is not feasible to be used for road safety and traffic
management applications. The results obtained from this study are shown in Fig. 7.
Time taken for establishing a connection reduces the available time for communica-
tion with transceivers in motion and should be taken into consideration while choosing
a technology for V2 V communication [4, 5, 36]. The time taken to establish a new con-
nection is 600 ms for Wi-Fi, 30 ms for ZigBee and 20 ms for DSRC [5]. Wi-Fi has the
largest time to establish a connection which reduces time available for communication
especially for relative speeds of 120 km/hr. Thus, Wi-Fi is not suitable for V2 V com-
munication [5].
The 2.4  GHz and 5.4  GHz unlicensed spectrum bands are used by Wi-Fi which is
prone to interference. This reduces the performance of Wi-Fi in practical scenarios as
packet loss increases due to collisions [36]. On the other hand, DSRC uses the carrier
frequency of 5.9 GHz. 75 MHz spectrum at 5.9 GHz is reserved in the US for DSRC for

Fig. 7  Average latency for DSRC and LTE for communication between moving transceivers (From [35])

13
 A. Thakur, R. Malekian

ITS applications that use V2 V and V2I communications [36]. Thus, interference related
packet loss is minimized in practical scenarios [36]. This makes DSRC a better choice
then Wi-Fi for V2I communications.
Amplitude of the signal reduces due to the presence of multipath effect introduced
by obstacles in the propagation environment. It was found in [36] that DSRC is more
robust against the multipath effect when compared to Wi-Fi. Thus, DSRC is more suit-
able for V2I communications especially in urban scenarios [31, 36, 37].
From the results obtained from the simulation in this study as well as in [5] and [35],
it was found that ZigBee, Wi-Fi and DSRC are capable for deployment in communica-
tion channels used in IoV systems. With the factors of time taken to establish a connec-
tion and success rate of the channel, DSRC and ZigBee can be used for V2 V communi-
cations which do not require a range larger than 100 m and high data rates [5, 35].
ZigBee offers low data rates and is incapable to offer successfully transfer of mes-
sages with high frequency of exchange events and so cannot be uses for V2I commu-
nication. With the factors of interference and transmission power of the radio, it was
found that DSRC is the most practical technology for V2I communication as it performs
better then Wi-Fi in urban environments where high path loss is introduced [37].

8 Conclusion

Road safety and traffic management are the major challenges faced in the transport
industry which directly impact quality of life. IoV solutions addresses these challenges
by using a wireless sensor network based approach that contain vehicular nodes, road-
side waypoint nodes and a central control unit [7].
This is achieved by using inter-vehicle communication and vehicle to infrastruc-
ture communication channels that are used for exchange of sensory data and generated
alerts. Road safety mechanisms in an IoV system include emergency braking warning,
abnormal driving condition alert system, collision avoidance systems and cooperative
awareness based precautionary alerting systems [5, 23]. Traffic condition assessment by
the use of connected vehicle technology is deployed in traffic management systems [20].
IoV solutions for traffic management use mechanisms such as optimized intersection
management and adaptive speed limits [1, 23, 38].
Effectiveness of the mechanisms used in an IoV solution [39–41] is dependent on
the performance and reliability of the communication links of the system in highly a
dynamic and mobile environment [23]. Thus, choosing a wireless communication tech-
nology is of prime importance.
From the simulations conducted in the identified platform of OMNeT ++, it was
found that ZigBee, Wi-Fi and DSRC can offer successful exchange of data with failure
rate of less than 1% for low frequency of communication events for a scenario with
vehicular nodes in motion with a speed of 120  km/hr. Whereas Wi-Fi and DSRC can
offer this performance at even higher frequencies of exchange events. The success rate
of the studied wireless communication technologies decrease with exchange of increas-
ing size and frequency of packets. However packets with a size of 300 Bytes can be
exchanged with high performance for a frequency of 10  Hz, which according to ETSI
requirements for ITS is the highest frequency of packet exchange.

13
Internet of Vehicles Communication Technologies for Traffic…

ZigBee and DSRC offer fast connection time and high performance for connection links
lasting for a short duration that require exchange of small packets. ZigBee and DSRC are
best suited for deployment in V2 V communication links.
Wi-Fi can transfer larger packets sizes when compared with DSRC with higher accu-
racy. But high transmission power of DSRC has robust performance in scenarios where
multipath effect is significant especially in urban scenarios [37]. Thus, DSRC is the more
practical wireless technology for V2I communication channels.
Connected vehicle technology that is deployed in IoV mechanisms use novel tech-
niques to enhance road safety and manage traffic. IoV solutions are robust as they offer
higher accuracy in assessment of traffic status and identification of events that can com-
promise road safety when compared to conventional systems [7]. Thus, IoV based systems
are effective in achieving the road safety and traffic management goals that are of primary
importance for intelligent transportation systems.

Acknowledgements The research project was supported by the National Research Foundation (NRF),
South Africa (grant numbers: IFR160118156967 and RDYR160404161474), T-Systems and Intervate (a
T-Systems Company), South Africa.

References
1. Barrachina, J., et al. (2015). A V2I-based real-time traffic density estimation system in urban scenar-
ios. Wireless Personal Communications, 83(1), 259–280.
2. Darwish, T., & Abu Bakar, K. (2015). Traffic density estimation in vehicular ad hoc networks: A
review. Ad Hoc Networks, 24, 337–351.
3. Zhu, Z., Loo, J., Chen, Y., Chai, K. K., & Zhang, T. (2016). Recent advances in connected vehicles via
information-centric networking. In IET international conference on intelligent and connected vehicles
(ICV 2016), Chongqing (pp. 1–8).
4. Hirose, K., Ishibashi, K., Yamao, Y., Hirayama,Y., & Sawada, M. (2015). Low-power V2 M commu-
nication system with fast network association capability. In 2015 IEEE 2nd world forum on internet of
things (WF-IoT), Milan (pp. 204–209).
5. Rawat, D. B., Bista, B. B., Yan, G., & Olariu, S. (2014). Vehicle to vehicle connectivity and communi-
cation framework for vehicular ad-hoc networks. In 2014 Eighth international conference on complex,
intelligent and software intensive systems, Birmingham (pp. 44–49).
6. Djajadi, A., & Putra, R. J. (2014). Inter-cars safety communication system based on Android smart-
phone. In 2014 IEEE conference on open systems (ICOS), Subang (pp. 12–17).
7. Qu, F., Wang, F. Y., & Yang, L. (2010). Intelligent transportation spaces: Vehicles, traffic, communica-
tions, and beyond. IEEE Communications Magazine, 48(11), 136–142.
8. Jeong, S., Baek, Y., & Son, S. H. (2016). A hybrid V2X system for safety-critical applications in
VANET. In 2016 IEEE 4th international conference on cyber-physical systems, networks, and applica-
tions (CPSNA), Nagoya (pp. 13–18).
9. Barrachina, J., Sanguesa, J., Fogue, M., Garrido, P., Martinez, F., Cano, J., et  al. (2013). V2Xd: A
vehicular density estimation system that combines V2 V and V2I communications. In 2013 IFIP Wire-
less Days (WD).
10. Eze, E. C., Zhang, S. J., Liu, E. J., & Eze, J. C. (2016). Advances in vehicular ad-hoc networks
(VANETs): Challenges and road-map for future development. International Journal of Automation
and Computing, 13(1), 1–18.
11. Siddiqi, K., Raza, A. D., & Muhammad, S. S. (2016). Visible light communication for V2 V intelligent
transport system. In 2016 International conference on broadband communications for next generation
networks and multimedia applications (CoBCom), Graz (pp. 1–4).
12. Chaqfeh, M., Lakas, A., & Jawhar, I. (2014). A survey on data dissemination in vehicular ad hoc net-
works. Vehicular Commununication, 1(4), 214–225.
13. Thakur, A., Malekian, R., & Bogatinoska, D. C. (2017). Internet of things based solutions for road
safety and traffic management in intelligent transportation systems. In D. Trajanov & V. Bakeva (Eds.)
ICT Innovations 2017: Data-Driven Innovation. 9th International Conference, ICT Innovations 2017,

13
 A. Thakur, R. Malekian

Skopje, Macedonia, September 18–23, 2017, Proceedings (pp. 47–56). Cham: Springer International
Publishing.
14. Hannan, M., Habib, S., Javadi, M., Samad, S., Muad, A., & Hussain, A. (2013). Inter-vehicle wire-
less communications technologies, issues and challenges. Information Technology Journal, 12(4),
558–568.
15. Siddiqi, K., Raza A. D., & Muhammad, S. S. (2016). Visible light communication for V2 V intelligent
transport system. In 2016 international conference on broadband communications for next generation
networks and multimedia applications (CoBCom), Graz (pp 1–4).
16. Yang, F., Wang, S., Li, J., Liu, Z., & Sun, Q. (2014). An overview of internet of vehicles. China Com-
munications, 11(10), 1–15.
17. Wan, N., Vahidi, A., & Luckow, A. (2016). Optimal speed advisory for connected vehicles in arterial
roads and the impact on mixed traffic. Transportation Research Part C: Emerging Technologies, 69,
548–563.
18. Du, Z., HomChaudhuri, B., & Pisu, P. (2017). Coordination strategy for vehicles passing multiple sig-
nalized intersections: A connected vehicle penetration rate study. In 2017 American control conference
(ACC).
19. Milanés, V., Villagrá, J., Godoy, J., Simó, J., Pérez, J., & Onieva, E. (2012). An intelligent V2I-based
traffic management system. IEEE Transactions on Intelligent Transportation System., 13(1), 49–58.
20. Khan, S. M., Dey, K. C., & Chowdhury, M. (2017). RealTime traffic state estimation with connected
vehicles. In IEEE transactions on intelligent transportation system. (pp. 1–13), July 2017.
21. Sanguesa, J., Fogue, M., Garrido, P., Martinez, F., Cano, J., Calafate, C., et al. (2013). An infrastruc-
tureless approach to estimate vehicular density in urban environments. Sensors (Switzerland), 13(2),
2399–2418.
22. Sanguesa, J. A., Barrachina, J., Fogue, M., Garrido, P., Martinez, F., Cano, J., et  al. (2015). Sens-
ing traffic density combining V2V and V2I wireless communications. Sensors (Switzerland), 15(12),
31794–31810.
23. ETSI, “ETSI TR 102 638 V1.1.1 (2009-06): Intelligent transport systems (ITS); vehicular communica-
tions; basic set of applications; definitions,” ETSI, Sophia (France), vol. 1, pp. 1–81, 2009.
24. Rehman Khan, A., Bilal, S. M., & Othman, M. (2012). A performance comparison of open source
network simulators for wireless networks. In 2012 IEEE international conference on control system,
computing and engineering (pp. 34–38).
25. Christhu, M., & Marium Chacko, N. (2013). A comprehensive overview on different network simula-
tors. International Journal of Engineering and Technology., 5(1), 325–332.
26. Minakov, I., Passerone, R., Rizzardi, A., & Sicari, S. (2016). A comparative study of recent wireless
sensor network simulators. ACM Transactions on Networking, 12(3), 20:1–20:39.
27. Arvind, T. (2016). A comparative study of various network simulation tools. International Journal of
Computer Science & Engineering Technology, 7(8), 374–398.
28. Rajaram, M., Kougianos, E., Mohanty, S., & Choppali, U. (2016). Wireless sensor network simulation
frameworks: A tutorial review: MATLAB/Simulink bests the rest. IEEE Consumer Electronics Maga-
zine, 5(2), 63–69.
29. Minakov, I., Passerone, R., Rizzardi, A., & Sicari, S. (2016). A comparative study of recent wireless
sensor network simulators. ACM Transactions on Sensor Networks, 12(3), 1–39.
30. Sommer, C., Härri, J., Hrizi, F., Schünemann, B., & Dressler, F. (2015). Simulation Tools and Tech-
niques for Vehicular Communications and Applications. In C. Campolo, A. Molinaro, & R. Scopigno
(Eds.), Vehicular ad hoc Networks: standards, solutions, and research (pp. 365–392). Cham: Springer
International Publishing.
31. Katsaros, K., & Dianati, M. (2016). A conceptual 5G vehicular networking architecture. In W. Xiang
& K. Zheng (Eds.), 5G mobile communications (pp. 593–623). Berlin: Springer.
32. Kurose, J., & Ross, K. (2013). Computer networking: A top-down approach (6th ed., pp. 25–70). Bos-
ton: Pearson.
3 3. Proakis, J., & Salehi, M. (2013). Digital communications (5th ed.). Boston: McGraw-Hill.
34. Bauza, R., Gozalvez, J. & Sanchez-Soriano, J. (2010). Road traffic congestion detection through coop-
erative vehicle-to-vehicle communications. In IEEE local computer network conference, Denver (pp.
606–612).
35. Dey, K., Rayamajhi, A., Chowdhury, M., Bhavsar, P., & Martin, J. (2016). Vehicle-to-vehicle (V2 V)
and vehicle-to-infrastructure (V2I) communication in a heterogeneous wireless network—Performance
evaluation. Transportation Research Part C: Emerging Technologies, 68(4), 168–184.
36. Lin, W.-Y., Li, M.-W., Lan, K.-C., & Hsu, C.-H. (2012). A comparison of 802.11a and 802.11p for
V-to-I communication: A measurement study. In X. Zhang & D. Qiao (Eds.) Quality, reliability,
security and robustness in heterogeneous networks: 7th international conference on heterogeneous

13
Internet of Vehicles Communication Technologies for Traffic…

networking for quality, reliability, security and robustness, QShine 2010, and dedicated short range
communications workshop, DSRC 2010, Houston, TX, USA, November 17–19, 2010, revised selected
papers (pp. 559–570). Berlin: Springer.
37. Gozalvez, J., Sepulcre, M., & Bauza, R. (2012). IEEE 802.11p vehicle to infrastructure communica-
tions in urban environments. IEEE Communications Magazine, 50(5), 176–183.
38. Zilu, L. & Wakahara, Y. (2013). City traffic prediction based on real-time traffic information for intel-
ligent transport systems. In ITS telecommunications (ITST), 2013 13th international conference (pp.
378–383).
39. Malekian, R., Kavishe, A. F., Maharaj, B. T., Gupta, P. K., Singh, G., & Waschefort, H. (2016). Smart
vehicle navigation system using hidden markov model and RFID technology. Wireless Personal Com-
munications, 90(4), 1717–1742.
40. Ye, N., Wang, Z., Malekian, R., Zhang, Y., Wang, R. (2015). A method of vehicle route prediction
based on social network analysis. Journal of Sensors, 2015, 1–9.
41. Ye, N., Zhang Y, Wang, R., Malekian, R. (2016). Vehicle trajectory prediction based on hidden markov
model. KSII Transactions on Internet & Information Systems, 10(7), 3150–3170.

Publisher’s Note  Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.

Arnav Thakur  received the B.E. degree as well as Honours degree in

computer engineering from the University of Pretoria, Pretoria, South
Africa, in 2015 and 2016 respectively. He is currently pursuing the
Masters studies in the Department of Electrical Electronic and Com-
puter Engineering. His research interests include intelligent transport
systems and advanced sensor networks.

Reza Malekian is with the Department of Computer Science and

Media Technology/Internet of Things and People Research Center,
Malmö University, Sweden and an Extraordinary Professor with the
Department of Electrical, Electronic and Computer Engineering at
the University of Pretoria. His research is focused on Internet of
Things, Sensors and smart cities. He is registered as a chartered engi-
neer with the Engineering Council of the UK. He is an associate edi-
tor for the IEEE Internet of Things Journal and IEEE Transactions on
Intelligent Transportation Systems.

13