Content uploaded by Vadym Hapanchak
Author content
All content in this area was uploaded by Vadym Hapanchak on Oct 28, 2023
Content may be subject to copyright.
University of Minho
School of Engineering
Department of Informatics
Vadym S. Hapanchak
Integration of different radio access technologies
for vehicular connectivity
December 2020
University of Minho
School of Engineering
Department of Informatics
Vadym S. Hapanchak
Integration of different radio access technologies
for vehicular connectivity
Doctoral dissertation planning document
Doctoral program in Advanced Engineering Systems for Industry
Dissertation supervised by
Prof. Dr. Ant´
onio Lu´ıs Duarte Costa
December 2020
Abstract
Vehicles have become an integral part of people’s lives in all countries and
brought several benefits to humanity. Given the advances in information
technology and communication, the concept of networked vehicles has re-
ceived immense attention from academia and industry worldwide. It allows
vehicles to exchange information among each other and with the infrastruc-
ture using appropriate technology. This feature can help prevent accidents,
improve traffic effectiveness, road safety and is referred to as Vehicle to Ev-
erything (V2X) communication. Recently, traditional V2X technologies have
been evolving to the Internet of Vehicles (IoV) to support the increasing de-
mands on emerging vehicular applications, such as advanced driving, Intel-
ligent Transportation Systems (ITS) and autonomous vehicles. Technology
used for vehicle communication should be designed to offer the solution with
a high degree of reliability and availability. It is a big technical challenge
due to high vehicle mobility, thus exploiting single wireless network is not
sufficient to support requirements of all V2X services. Vehicle manufacturers
intend to increase connectivity capabilities of the communication equipment
providing access for multiple wireless interfaces. Next generation vehicular
networks will be highly dynamic and exploit various wireless technologies
including Wi-Fi and cellular networks to form the IoV. This way, a variety
and coexistence of wireless networks form a heterogeneous vehicular commu-
nication environment. Therefore, an efficient integration of different Radio
Access Technology (RAT)’s is one of the most addressed research issues in
the context of heterogeneous networks. In the same context, seamless con-
nectivity (i.e. continuous connectivity) is a huge challenge in V2X, caused
by the extremely dynamic network topology and technology diversity. The
integration of different wireless technologies offers many potential benefits for
vehicle communication system, for instance, high data rates, low latency, and
extended communication range. Providing continuous connectivity in such
heterogeneous environments with highly variable number of mobile nodes is
a huge challenge. Switching across different networks disrupts ongoing con-
nections, resulting in high packet losses and delays, and consequently affects
i
service performance. Therefore, the implementation of a seamless V2X solu-
tion is one of the top priority objectives. Information presented in this thesis-
planning document describes the basic characteristics of V2X communication
and reviews some related studies in the literature that explore heterogeneous
wireless networks. With this document, we create a solid background to start
design a new V2X solution that combines multiple technologies to provide
best vehicle connectivity at all times. We have highlighted the key techni-
cal issues and mentioned the opportunities for future research toward the
integration of multiple RAT’s in a heterogeneous environment.
ii
Contents
1 Introduction 1
1.1 Objectives............................. 2
1.2 Document Structure . . . . . . . . . . . . . . . . . . . . . . . 3
2 Overview of V2X technologies 4
2.1 Architecture............................ 5
2.2 V2XApplications......................... 8
2.3 Evolution of V2X standards . . . . . . . . . . . . . . . . . . . 9
2.3.1 DSRC/IEEE802.11p . . . . . . . . . . . . . . . . . . . 9
2.3.2 C-V2X........................... 11
2.3.3 Wi-Fi ........................... 13
2.3.4 IEEE802.11bd ...................... 15
2.3.5 5G-V2X .......................... 15
2.4 V2Xequipment.......................... 18
2.5 Heterogeneous communication . . . . . . . . . . . . . . . . . . 19
2.5.1 Handover in V2X . . . . . . . . . . . . . . . . . . . . . 20
2.5.2 Seamless handover . . . . . . . . . . . . . . . . . . . . 21
3 Challenges and opportunities 23
4 State of the art 25
4.1 Review of V2X domain . . . . . . . . . . . . . . . . . . . . . . 26
4.2 Standard Approaches . . . . . . . . . . . . . . . . . . . . . . . 27
4.3 NetworkSelection......................... 28
4.4 IP-based solutions . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.5 Multipath Solutions . . . . . . . . . . . . . . . . . . . . . . . . 34
4.6 Software-defined networking . . . . . . . . . . . . . . . . . . . 37
4.7 Cognitiveradio .......................... 41
5 Work Plan 44
iii
List of Figures
2.1 Vehicles network architecture [1]................. 6
2.2 V2X communication types .................... 7
2.3 The WAVE protocol versus ITS-G5 stack ........... 10
2.4 C-V2X transmission Modes ................... 12
2.5 Evolution of V2X standards ................... 16
2.6 Ongoing 5G technology trends .................. 17
2.7 Examples of the V2X boards ................... 18
2.8 New generation V2X equipment supports multiple RAT . . . . 19
2.9 Horizontal and vertical handover in vehicular HetNet ..... 21
4.1 Radio interface selection function ................ 30
4.2 MPTCP position in protocol stack ............... 34
4.3 MPTCP in vehicular HetNet communication scenario . . . . 35
4.4 A topological architecture of software-defined vehicular HetNet 39
4.5 Software-defined radio approach ................. 43
5.1 Work Plan Timeline (divided by Quarters) ........... 47
v
Acronyms
3GPP 3rd Generation Partnership Project.
ADAS Advanced Driver Assistance System.
AESI Advanced Engineering Systems for Industry.
ANDSF Access Network Discovery and Selection Function.
API Application Programming Interface.
BTP Basic Transport Protocol.
C-V2X Cellular Vehicle to Everything.
CAN Controller Area Network.
CCU Communications Control Unit.
CR Cognitive Radio.
DCC Decentralized Congestion Control.
DSRC Dedicated Short-Range Communications.
ECU Electronic Control Unit.
ETSI European Telecommunications Standards Institute.
FMIPv6 Fast Handover Mobile IPv6.
GPS Global Positioning System.
HetNet Heterogeneous Networks.
vii
HHO Horizontal Handover.
HMI Human-Machine Interface.
IEEE Institute of Electrical and Electronics Engineers.
IoT Internet of Things.
IoV Internet of Vehicles.
IP Internet Protocol.
ITS Intelligent Transportation Systems.
LTE Long-Term Evolution.
MCDM Multi-Criteria Decision Making.
MEC Mobile Edge Computing.
MIH Media Independent Handover.
MIPv6 Mobile IPv6.
MPIP Multipath IP.
MPQUIC Multipath QUIC Protocol.
MPTCP Multipath TCP.
NFV Network Function Virtualization.
NLOS Non-Line of Sight.
OBU On-Board Unit.
OEM Original Equipment Manufacturer.
PMIPv6 Proxy Mobile IPv6.
QoS Quality of Service.
QUIC Quick UDP Internet Connection.
RAT Radio Access Technology.
viii
RF Radio Frequency.
RSSI Received Signal Strength Indicator.
RSU Roadside Unit.
RTT Round-Trip Time.
SDN Software Defined Networking.
SDR Software-defined Radio approach.
TCP Transmission Control Protocol.
V2I Vehicle to Infrastructure.
V2N Vehicle to Network.
V2P Vehicle to Pedestrian.
V2S Vehicle to Sensor.
V2V Vehicle to Vehicle.
V2X Vehicle to Everything.
VANET Vehicular Ad Hoc Networks.
VHO Vertical Handover.
VRU Vulnerable Road User.
WLAN Wireless Local Area Network.
ix
Chapter 1
Introduction
The statistics show that the number of vehicles per habitant in Europe is
growing annually. V2X communication technology has emerged to enhance
traffic safety and transportation efficiency by empowering vehicles with the
ability to communicate and interact with each other and infrastructure. V2X
is a term that is typically used to describe technologies that allow vehicles
to send and receive data from any entity in the vehicular communication
environment. The main V2X objective is to ensure road safety, reduce fuel
consumption and offer the best level of driving comfort. It enables vari-
ous applications and services to be accessed, including safety, infotainment,
mobility, payment, and so on.
IoV is the next stage of vehicle communication evolution that can sup-
port a wider range of advanced ITS applications, such as autonomous vehi-
cles, remote and cooperative driving, real-time environmental perception and
control. IoV links various entities of transportation systems, such as pedestri-
ans, vehicles, road infrastructure, and cloud environments. IoV enables the
evolution from single-vehicle intelligence to connected vehicle intelligence,
increases vehicle perception, computation and communication capabilities,
which is beneficial for many ITS applications [2]. By sharing information
within the network, vehicles are able to solve transportation system issues
more effectively in a cooperative way. The IoV is going to offer increased
environmental perception to enable sensor-data sharing among vehicles and
infrastructures. These advanced applications require real-time network ac-
cess, extensive data transmission that are beyond the capability of single-
technology communication devices. Furthermore, IoV must assure vehicles
continuous connectivity and information exchange without having to worry
about transitions between and across Heterogeneous Networks (HetNet). An
IoV environment integrates human, vehicle and “things” that collaborate
with each other.
1
Modern vehicles are going to be connected through heterogeneous RAT
(multi-RAT) and they are going to exchange massive information with their
surrounding environment [3]. Different applications have different communi-
cation requirements. All of the wireless technologies adopted for V2X have
its advantages and limitations. They are characterized by capacity, coverage,
latency, data rate, mobility, spectral efficiency, and reliability. Improving
these parameters simultaneously in a single technology is not that simple,
and it may not deliver everything. There are many challenges associated
with single-RAT usage to fulfil IoV requirements such as high availability,
mobility support, reliability and network scalability. Therefore, the interop-
erability of different wireless technologies is still an active research area and
it lacks new mechanisms for an improvement of the services offered by the
V2X.
1.1 Objectives
This thesis proposal is aligned with the AESI doctoral program resulting
from the collaboration of the University of Minho and Bosch Car Multime-
dia, in which all activities are aimed at reaching application in real company
needs. “Easy Ride: Experience is Everything” - Sub-project P24 - V2X
Vehicle-to-Everything Communications is a collaborative R&D project dur-
ing which Vehicles On-Board Unit (OBU) is being developed and relative
ITS applications created and tested. OBU is communication equipment for
cars that support multiple network interfaces to provide connectivity in Het-
Net. Therefore, new algorithms and strategies on the interoperability of
different types of wireless standards within the same OBU are required to be
investigated. Moreover, these solutions should use efficiently technological
resources currently available. Within the framework of this research project,
it is expected to identify scenarios, use cases and other related data that will
be useful for the thesis work.
The focus of this doctoral work is the definition, evaluation and test of
an advanced communication technique based on multi-RAT capabilities of
vehicle OBU. It should be possible to take advantage of HetNet to improve
the overall communication system characteristics and performance. Besides
showing the benefits and main design aspects when employing multi-RAT
diversity, the goal is to quantify the reliability improvement achieved in co-
operative driving use cases. A prototype implementation of the resulting
system should be done as a proof of concept. It should be validated and
tested in a real world deployment using provided equipment; however, the
result should be independent of any specific hardware. The system should
2
also be tested in a simulation environment to evaluate solution performance
in a large-scale scenario. Simulation can enable a better understanding of its
adoption in a real-world implementation. A survey of simulation tools should
be done to select the framework with realistic network communication and
vehicle dynamics better suited for this case. During this thesis works, sce-
narios and use cases of the proposed methods should be identified. The final
goal is to derive, validate and test the new communication technique for
vehicular HetNet and to validate the implemented solution concerning its
characteristics such as latency, throughput, packet delivery ratio, reliability.
1.2 Document Structure
The remaining of this document is organized as follows: Chapter 1 presents
objectives that are intended to be achieved during the PhD research work.
Chapter 2 makes a brief overview of the vehicle’s connectivity, describing the
architecture, applications and radio access technologies. Also, an introduc-
tion to IoV environment is done and the concept of network heterogeneity is
presented. The use of multiple networks with different characteristics repre-
sents several challenges for moving nodes, like vehicles. Therefore, Chapter
3 summarizes the research questions addressed in this PhD. Then, Chapter
4 describes the current state of the art of technologies related to the work
developed in this thesis, including standard solutions, characteristics and re-
quirements for vehicle communications in a heterogeneous environment with
research initiatives and possible research directions. Finally, Chapter 5 de-
scribes the work plan and the tasks list to be followed. Time should be
allocated for each one of the tasks to be carried on. This chapter finishes
with a Gantt chart that presents the elaborated plan.
3
Chapter 2
Overview of V2X technologies
Vehicular communication is a major component of the ITS framework that
aims to contribute towards safer and efficient road traffic. In the last few
years, increasing attention has been provided to research the Internet of
Vehicles: the evolution of conventional Vehicular Ad Hoc Networks to global
heterogeneous vehicular networks. VANET focus remains on the aspect of
spontaneous networking through short-range information exchange, much
less on the use of cellular networks. VANETs can use any wireless technology
as their basis, however, communication standards, such as the Dedicated
Short-Range Communications (DSRC) [4] and Cellular Vehicle to Everything
(C-V2X) [5], designed specifically for automotive use. The main motivation
for deploying VANET at an early stage is to enable road safety applications,
such as collision warning and emergency event informing, by direct messaging
among vehicles and roadside infrastructure. However, neither DSRC nor C-
V2X alone cannot meet requirements for future IoV applications, such as
autonomous vehicles. For example, the DSRC standard has been designed
explicitly for vehicular communications and provides seamless handover by
design, but it has a low market penetration, requires a specific infrastructure
installation and suffers in terms of data rate.
To support the recently emerged needs of better communication and in-
terconnectivity between vehicles, conventional V2X technologies are evolving
to the IoV, to achieve the vision of ‘from smartphone to smart car’ [6]. IoV
provides wide and real-time network access for emerging automotive appli-
cations, including autonomous driving, intelligent traffic control, and collab-
orative environmental perception. These applications require extensive data
transmission and information exchange that is beyond the capability of con-
ventional connected vehicles [7]. IoV allows real-time communication among
different network entities, such as vehicles, road infrastructure, pedestrians,
parking lots and so on. This creates a network with intelligent devices as
4
participants where all vehicles and infrastructure systems are interconnected
with each other. Therefore, IoV is so important for the evolution of automo-
tive. The development of IoV passes through the capacity to interconnect
humans, vehicles and multi-level collaboration systems by sensors and mo-
bile devices into a global network. Every entity in the IoV environment can
consume/produce data and interact with the vehicle. Due to this internet-
working environment in IoV, different wireless technologies can be used to
establish connections.
This chapter overview the vehicle network architecture, briefly describes
applications for connected vehicles, main actual wireless technologies, dif-
ferent V2X communication types, the underlying system of the vehicular
HetNet and equipment for vehicle communication.
2.1 Architecture
To fully describe what vehicular connectivity is, the V2X network architec-
ture should be presented. Figure 2.1 illustrates the generic V2X network
architecture consisting of four distinct layers: In-vehicle, Ad hoc, Infrastruc-
ture, and Service domain.
In-vehicle domain is an intra-vehicle system including the Communi-
cations Control Unit (CCU), the On-Board Unit (OBU) and the Human-
Machine Interface (HMI). Vehicles can support multiple network interfaces
to provide connectivity in a HetNet, e.g., a combination of DSRC and cellu-
lar networks. CCU ensures connectivity to all networks needed and contains
transceiver modules with radio frequency antenna for each communication
interface. Also, it enables integration to different in-vehicle sensors (GPS,
RADAR, Camera). It provides data derived from internal communication
and external communication to the OBU. The OBU is equipped with hard-
ware (processing, memory, storage and communication capabilities) and soft-
ware to run applications. The primary roles of OBU are wireless radio access,
network congestion control, traffic routing, reliable message transfer, data se-
curity and IP mobility. Furthermore, OBU collects environmental data for
detecting events, driving patterns, and situations. This computing system is
designed to support a wide variety of ITS services and provides an interface
to HMI for driver interaction. The HMI is a user interface that facilitates the
interaction between the driver and vehicle; enables drivers to use the capa-
bilities of OBU and provides relevant information. This information can be
presented to the driver via display screen or voice to provide better assistance
and interaction capabilities.
In the ad hoc domain, the wireless network is created spontaneously to
5
Figure 2.1: Vehicles network architecture [1]
provide vehicle communication (i.e., VANET). The network does not rely on
pre-existing infrastructure, so the topology varies rapidly and unpredictably.
Each mobile node connected to the others forms a partially connected mesh
network (a graph of arbitrary size). VANET nodes (cars) can move ran-
domly and organise themselves arbitrarily. It is a key part of ITS framework
that allows V2X communication. Roadside Unit (RSU) is a network device
placed at dedicated locations along roads and supports at least one ad hoc
technology (e.g. DSRC) to interact with cars. Vehicle can be seen as a mobile
node of ad hoc network and RSU as a static node similarly. Communication
in VANET can be one-hop or multi-hop (i.e., the dissemination of messages
over vast distances). The presence of RSU can help extend the range of
communication and take part in forwarding messages to other vehicles.
Infrastructure domain includes the roadside wireless infrastructure (e.g.
RSU, LTE Base Stations (eNB), WiFi Hotspots) and the backbone wired
network. These wired and wireless components can be deployed by ITS
authorities or service providers, allowing vehicles access to the infrastruc-
ture network. Infrastructure provides connectivity between vehicles and the
6
service domain. Service domain is the top layer of the architecture that pro-
vides services to the vehicles through infrastructure. Numerous examples
of possible services exist, e.g., traffic-related services, internet connection,
infotainment services, manufacturer based, and other cloud-based services.
Figure 2.2: V2X communication types
In addition, vehicle network architecture distinguishes different communi-
cation types, for instance Vehicle to Vehicle (V2V), Vehicle to Infrastructure
(V2I), Vehicle to Pedestrian (V2P), Vehicle to Network (V2N), Vehicle to
Sensor, etc. Term Vehicle to Everything represent a generalisation for all
these specific modes and consists in the data transfer from a vehicle to any
entity or vice-versa (Figure 2.2). Communication with surrounding environ-
mental IoT devices shall be possible when new V2X technologies be incor-
porated into the vehicle. Thus, IoV is an ecosystem that links various ITS
elements, such as automotive, vulnerable road users (VRU, e.g. pedestri-
ans), road infrastructure equipment and cloud environments, and guarantees
cooperation between them.
V2V is a direct communication between cars. The primary purpose of
V2V wireless data transmission is to prevent possible accidents, allowing ve-
hicles in transit to share information (e.g. position, speed) within an ad hoc
mesh network. If two vehicles are in their communication range, they can
exchange messages directly; if not, they can use multi-hop communication
to send data, by using a dedicated routing protocol, forwarding data from
7
one node to another until it reaches the destination point. V2V type typi-
cally used to support cooperative application and to exchange time critical
information with nearby vehicles. This could be, for example, an emergency
warning system or cooperative adaptive cruise control.
V2I is a communication between the vehicle and transport infrastructure
(e.g. RSU or LTE Base Station), and vice-versa, mainly for data gather-
ing applications. Vehicle can use cellular and road infrastructure gateways
to access the services on the Internet, thereby integrating car devices into
the IoV. An example of road infrastructure could be the traffic lights in an
intersection with a RSU connected to backbone network. Automotive com-
munication devices could download, for example, current weather forecasts
(e.g., warning about poor road conditions), traffic information or infotain-
ment data. Alternatively, traffic supervision and management systems can
use the data collected from the road infrastructure and vehicles to control
traffic flows.
2.2 V2X Applications
The primary goal of vehicular communication is to provide safety and com-
fort for passengers. Based on data from various studies and surveys, nu-
merous applications of connected vehicular technologies can be identified [8].
VANET potential applications are targeted to on-road safety, transport effi-
ciency and infotainment services on the move, but the key features of V2X
focus on ultra-reliable and low latency communication for safety-critical use
cases. Different applications have different communication requirements in
terms of latency, reliability and throughput. Applications implemented on
top of VANET may use single-hop or multi-hop connections. IoV introduces
several additional applications, which are bringing fundamental changes to
transportation system management, such as autonomous driving, intelligent
traffic control, crash response, convenient services, social behaviours [9].
Most of the vehicular applications fall into two broad categories: safety
and non-safety. A safety application requires an extremely low latency and
a secure network environment. It refers to applications involving personal
safety, such as collision avoidance, speeding warnings, pre-crash sense or
lane changing. They provide knowledge about the vehicle direct environ-
ment, which helps prevent road accidents. Safety applications have strict
communication requirements in terms of maximum allowable latency and
range. Non-safety applications enable passengers to access all other services
related to mobility, environment, infotainment, social networking, mainte-
nance, payment services, etc. For instance, a congestion warning and optimal
8
speed advisory application provide information to vehicles in transit that can
help reduce fuel consumption.
Vehicular applications should be rigorously tested to ensure the reliabil-
ity and maturity before introduction into the market in mass. At present,
IoV is still in the exploratory stage, and the problems of traffic safety and
information security brought by ITS applications have not yet been fully
evaluated.
2.3 Evolution of V2X standards
V2X communication is a solution that is associated with the set of standards
that will enable vehicles to interact with other road users and infrastruc-
ture. Currently, there are two major standardized technologies considered
to be used for V2X: Dedicated Short Range Communication (DSRC) and
Cellular-V2X (C-V2X). Both of them operate in the 5.9 GHz band but C-
V2X also in the cellular operator’s licensed spectrum. Although DSRC is the
first standard for vehicular communications, C-V2X is also relevant for the
V2X because in some cases they offer better performance than the DSRC-
based networks. These two technologies were enhanced in many aspects to
accommodate exchanges of rapidly changing dynamic information (position,
speed, driving direction, etc.) and future advanced V2X services (automatic
driving, vehicle platooning, sensor sharing, etc.). At present, there is no
massive integration on the market of any of these technologies and they are
still in the exploratory stage. This section outlines mainstream technologies
of wireless vehicle communication, including the most advanced 5G, WiFi 6
and IEEE 802.11bd.
2.3.1 DSRC/IEEE802.11p
DSRC is a first standard and the primary wireless technology enabling V2X
communication. It has been 20 years since the first licensed spectrum was
allocated for DSRC to provide vehicle connectivity. The main motivation for
deploying V2X based on DSRC at an early stage was to enable “day one” road
safety applications, such as frontal collision warnings, blind spot warnings,
and intersection motion assistance [7]. In 2008, the European Telecommuni-
cations Standards Institute (ETSI) allocated 70 MHz of spectrum in the 5.8
GHz band for DSRC automotive applications. At the physical layer (PHY)
and the lower medium access control layer (MAC), DSRC based on the IEEE
802.11p protocol, that is used nowadays exclusively in vehicle communica-
tion. IEEE 802.11p operates without network coverage in a fully distributed
9
manner. It simplifies associated authentication processes and allows effec-
tive and almost immediate data transmission, enabling vehicles to broadcast
relevant information. The protocol can also defer authentication, encryp-
tion and full identification to higher level protocols. Therefore, vehicles and
infrastructure can begin exchanging essential data on speed and position
within tens of milliseconds of detecting each other, which is obviously ideal
for applications such as collision avoidance. The network architecture and
security protocols are defined in the upper layer amendments: ETSI ITS-G5
(in Europe) and IEEE1609 WAVE (in the US) (Figure 2.3). Both variations
support TCP and UDP over IPv6, but for time-critical features they each
have their own specialised low-overhead messaging standard. WAVE uses the
wave short message protocol (WSMP), which is defined as part of IEEE 1609.
However, ITS-G5 employs a multi-hop routing system, the Basic Transport
Protocol (BTP) over the GeoNetworking service. GeoNetworking is a ge-
ographically aware routing technique, which establishes an ad-hoc network
that is efficiently arranged according to the physical locations of nodes. The
ad hoc nature of DSRC technology brings advantages such as independent
operation from cellular infrastructure, ubiquitous connectivity, low control
overhead and potentially low delays. Also, WLAN nature of the technology
brings benefits such as the easier wireless station deployment. However, due
to the random medium access schemes, the performance of ad hoc technolo-
gies degrades under channel congestion, resulting in packet loss and increased
update times. Furthermore, despite significant research done on 802.11p, it
suffers from scalability issues, limited coverage area, and unbounded delay
[10].
Figure 2.3: The WAVE protocol versus ITS-G5 stack
DSRC supports Orthogonal Frequency Division Multiplexing (OFDM)
technique for data multiplexing with a raw bandwidth of 10 MHz. OFDM
10
divides the input data stream into parallel bit streams and maps individual
bit streams onto overlapping sub-carriers. DSRC gives medium data transfer
rate ranging from 6 Mbps to 27 Mbps, half duplex, at short radio trans-
mission distance (approximately 300 - 1000 m), low latency, and supports a
vehicle speed up to 200 km/h. At the MAC layer, carrier sensing multiple
access with collision avoidance (CSMA/CA) is adopted. A node needing to
transmit senses the medium to check if it is idle, and a mechanism based
on random backoff is performed to reduce the probability of collisions. The
CSMA/CA mechanism introduces significant overhead, especially with short
packets and high-level modulations. In addition, communication is collision
affected when the channel occupation raises 50%–60% [11], thus mechanisms
to avoid overloads are necessary. This brought ETSI through Decentral-
ized Congestion Control (DCC) to add algorithms specifically addressing
congestion control. Therefore, research communities still see room for im-
provements to the current DSRC solutions at PHY and MAC layer. IEEE
802.11px is the enhanced version of the IEEE 802.11p protocol that brings
the new channel-coding mechanism to improve the packet delivery rate, and
provides MIMO antenna capabilities. Due to the limited bandwidth and com-
munication range, DSRC cannot provide high data rate transmissions, and
the connection is frequently interrupted with the Non-Line of Sight (NLOS)
conditions and highly dynamic network topology. Recent studies show that
DSRC can support end-to-end transmission latency of 50–100-ms, which can
be sufficient for day one application [7]. However, for some advanced IoV
applications such as remote driving, DSRC is not usable as they require
more stringent transmission latency. DSRC was a first standard adopted for
V2X communication, but after 20 years from publication, it is difficult to
say that DSRC is successful or even common in the automotive market. The
main issues to solve are connection interruption, low data rate and coverage
limitation.
2.3.2 C-V2X
Cellular Vehicle to Everything (C-V2X)) is a vehicle communication technol-
ogy developed based on the cellular systems. C-V2X is the first LTE technol-
ogy expansion that focuses on automotive services introduced by 3GPP [8].
C-V2X is a complex IoV service network for vehicle communication with co-
ordinative interaction and cognitive support computing [2]. It is based on the
evolution of LTE-A mobile communication technology for automotive needs,
with high data rate, high spectral efficiency and controlled QoS. The C-V2X
(Release 14/15) provides low latency (10–20 ms), high reliability (95%) and
high throughput (30 Mbps) for V2X communications by leveraging and en-
11
(a) LTE-Uu mode (b) PC5 (side link) mode
Figure 2.4: C-V2X transmission Modes
hancing current cellular systems with transmission range up to 30 km in the
radio network.
C-V2X employs two complementary transmission modes to support the
ITS services. V2X communication can be implemented by forwarding data
through the cellular network (LTE-Uu mode) or by direct communications
provided by PC5 (sidelink), e.g., when the vehicle is out of the network cov-
erage (Figure 2.4). The sidelink mode is similar to the DSRC, enabling direct
communication between vehicles with the low transfer latency, and does not
require any mobile subscription (SIM card) or mobile network infrastructure.
Authors [12] provide essential knowledge of sidelink transmissions including a
new control channel design, channel sensing procedure, physical layer struc-
ture, resource allocation mechanisms, synchronization and QoS management.
For V2X direct communications, C-V2X uses the same 5.9 GHz spectrum as
DSRC. However, C-V2X can also operate in traditional cellular bands to
communicate with the mobile network (i.e., V2N mode).
The resource allocation algorithm plays a crucial role to optimise the
spatial reuse on short-range Side link communication. Unlike DSRC, V2X
communications over PC5 interface can use either network-scheduled Mode
3 (i.e., centralised) or autonomous selection Mode 4 (i.e., distributed). Mode
3 algorithms are not defined in the specifications and their implementation
is left to the operators. By allowing direct communication among vehicles
within proximity areas bypassing the infrastructure, V2X traffic can be of-
floaded from the cellular network core. Thus, better network throughput,
energy consumption, spectrum utilisation, and significantly lower delay per-
formance can be achieved.
V2X communication over the LTE-Uu interface is supported only when
the vehicle is inside LTE network coverage. In contrast to DSRC, trans-
12
missions in this communication mode are scheduled by a networks eNB (i.e.
centralised), thus collisions and interference can be controlled. Furthermore,
the scheduler can provide QoS guarantees to different applications by admis-
sion control and radio resource allocation, which is a major advantage over
the CSMA/CA access scheme in 802.11p.
The main C-V2X challenge is to have dedicated spectrum needed for
safety applications since LTE networks are easily overloaded with an in-
creasing number of users, causing serious latency during peak hours. Other
challenges for instance, managing resources in high-density scenarios, control
and channel estimation overhead (especially with short payload), resource
granularity and channel coding.
Many researchers have concluded that C-V2X will be a strong competitive
technology against 802.11p and will show steady growth in future [13, 2].
Some papers promote C-V2X as a more efficient technology, supported by
a strong existing infrastructure and a clear road-map for future evolution
[11]. C-V2X provides an evolution road from LTE to 5G and can technically
coexist with 802.11p-based radio access in adjacent channels. In addition, it
can support the upper protocol layers of DSRC – effectively replacing only the
IEEE 802.11p MAC and PHY radio layer with 3GPP-specified equivalents,
while allowing the specialised higher-level application to run as normal.
The C-V2X standardisation process in Europe is speeded up in 2018
and the related C-V2X standards work has been completed. The protocol
stack has been defined to provide the availability of the C-V2X. 5GAA (the
5G Automotive Association), as a global and cross-industry organisation,
has presented the clear position of supporting C-V2X as feasible solutions
for the mobility and transportation services to support connected cars and
road safety applications [3]. More than 110 companies from the automotive,
telecommunications, and IT industries have now joined 5GAA. The Day-one
function profile and system profile have been completed in 2018. In 2019,
13 Chinese OEMs jointly made an announcement that they would support
volume production of C-V2X vehicles from the second half of 2020 to the first
half of 2021. The European 5GCAR project [14] investigated and proved the
added value of C-V2X for connected cars, studying the domain from multiple
perspectives.
2.3.3 Wi-Fi
Wi-Fi-based V2X has attracted much attention due to its unlicensed spec-
trum advantage and high performance even on roads, bringing the success
of Wi-Fi to vehicle networks. Many studies have proved the utility of Wi-Fi
networks to enrich the information exchange methods among vehicles and
13
infrastructure [8]. Wi-Fi is seen as a complementary and low-cost solution to
provide V2I connectivity and can be used for many applications (e.g., vehi-
cle data offloading, content caching). However, those implementations have
reduced effective communication ranges of only a few hundred meters, so,
to maintain connectivity in motion, vehicles reconnect from one AP to an-
other. In a real automotive environment, this could easily result in frequent
handovers. To enable seamless AP switch when the vehicle drives through
a series of roadside Wi-Fi networks, IEEE 802.11r amendment (also called
fast roaming) is issued to permit continuous connectivity for wireless devices
in motion, with fast and seamless handovers from one AP to another. IEEE
802.11r standard specifies fast Basic Service Set (BSS) transitions between
AP’s by redefining the security key negotiation protocol, allowing both the
negotiation and requests for wireless resources to occur in parallel.
The new Hotspot 2.0 specification is an approach to public access Wi-
Fi. It is a Wi-Fi access framework, including the automatic association,
secure communication, and better interworking with core networks [8]. The
goal of Hotspot 2.0 networks is to provide cellular-style “roaming” for Wi-Fi
networks, and on a technical level, it’s based on the 802.11u Wi-Fi standard
published in 2011. As a vehicle moves along the road, the mobile device
will connect to available public hotspots automatically and join a subscriber
service whenever it enters a Hotspot 2.0 area. The technology is still new,
and most of the Wi-Fi APs is not Hotspot 2.0-enabled, thus it will take some
time to gain the wide coverage necessary to replace the older networks.
To apply Wi-Fi technology to vehicular networks, there are some issues
because it is not designed for high mobility context. The connection to
roadside Wi-Fi networks would be frequently interrupted since Wi-Fi has a
small coverage range (100–300 m). When APs are densely deployed in urban
areas, there will be frequent handoff that can interrupt ongoing communi-
cations as well. Security, QoS support, seamless connectivity, Doppler shift,
latency spread and multipath effects are some of the most significant chal-
lenges towards Wi-Fi V2X implementation. However, Wi-Fi networks keep
improving in terms of higher link rate (up to 1 Gbps), mobility support with
the low deploying cost. It’s easy to build a roadside AP based on inexpensive
hardware and open-source software. Thus, considering reduced deployment
price and network performance, Wi-Fi should be widely accepted for IoV.
Recently, new solutions based on the latest WiFi 6 standard were introduced
optimised for vehicle operation [15]. It is a most advanced Wi-Fi solution
designed to offer fast, secure and efficient Wi-Fi connectivity, to meet V2X
demands for greater robustness and reduced latency when operating in con-
gested and dense environments. Wi-Fi 6 is designed to deliver efficient Wi-Fi
connectivity throughout the vehicle, enables better use of frequency spectra
14
and offers users a bandwidth of up to 1.8 Gbps. Full multiple-input mul-
tiple output (MIMO) capability designed to extend the range at high data
rates for connecting to external access points for automotive services, such
as vehicle diagnostics, software updates and automatic check-ins.
2.3.4 IEEE 802.11bd
It has been shown that DSRC performance is satisfactory for most V2X
applications that require end-to-end latency to be less than 100 msec [9].
However, if the density of vehicles increases, performance rapidly drops due to
packet collisions because of multiple simultaneous transmissions and packet
collisions. The DCC mechanism, implemented in ITS-G5 standard, purposes
to improve poor scalability of DSRC.
With the aim to enhance the existing 802.11p technology with advanced
PHY and MAC techniques introduced in the newest Wi-Fi standards, a new
IEEE study group for next-generation V2X has been announced. After an
initial study, the IEEE 802.11bd Task Group was created in January 2019.
The primary design objectives of 802.11bd include supporting high mobil-
ity (up to 500 km/h), high data rate and increased communication range
(up to 2 km). Emphasis was posed to modifications at the PHY layer, in
particular, to MIMO techniques, improved channel coding, and better pilots
placing. Some noteworthy features of V2X standards are outlined in Fig-
ure 2.5. Additionally, 802.11bd should guarantee coexistence and backward
compatibility with 802.11p. The study group is in the initial stage investi-
gating the usage of more advanced PHY layer technologies in amendments
after 802.11p and is expected to be completed in 2022.
The paper in [9] survey the latest developments in the standardisation
of 802.11bd and 5G New Radio (NR) V2X, making a brief description of
the technologies itself, outline their objectives and features, as the descrip-
tion of key mechanisms that enable these features. While both, the IEEE
802.11bd and 5G, are in their initial stages of development, this work de-
scribes their preliminary performance projections and compares the two evo-
lutionary RATs with their respective predecessors.
2.3.5 5G-V2X
Fifth-Generation Network (5G) is one of the most recent and promising tech-
nologies, now gaining the momentum on IoV research field, guaranteeing fast
broadband connections and security. 5G has been defined to support different
services and use cases (e.g., mobile broadband, industry 4.0 and connected
vehicles). It is expected that 5G networks will transmit a great volume of
15
Figure 2.5: Evolution of V2X standards
data (20 Gbps), provide ultra-low latency (less than 1 ms), high mobility (500
km/h) with sub-meter positioning accuracy. That makes 5G network tech-
nology very appealing for automotive use cases. 5G shows promising signs
in terms of relative improvements over its predecessor technology. 5G can
be used to support V2X communications and applications on autonomous
vehicles in a much better way than actual RATs. Unlike LTE, it does not
refer to a particular standard, but rather a collection of requirements and
new technologies that are seen as the next step for mobile cellular networks.
5G is expected to have flexible design to support the stringent requirements
of V2X services, i.e. extreme low latency, better coverage, high system ca-
pacity, high reliability. The future 5G network is expected to come with more
flexible and flat IP architecture and supported with V2X services closer to
the users. Therefore, 5G is seen as a main network technology to support
future IoV with secure and fast connection [14].
5G system requires the introduction of some fundamentally new tech-
nologies and paradigms to complement ongoing evolutionary trends. The
5G-based cellular systems are going to have the mmWave communication
that operates in the spectrum between 30 GHz and 300 GHz. Based on
the 5G interface design, the new physical layer structure of Side link sig-
nals, channels, bandwidth parts, and resource pools is proposed to support
more transmission types (e.g., group cast). With the 5G novel core net-
work architecture and technical features such as Mobile Edge Computing
16
Figure 2.6: Ongoing 5G technology trends
(MEC), Network Function Virtualization (NFV), Software Defined Network-
ing (SDN), can be realized to satisfy different service requirements. Several
ongoing 5G trends are identified in Figure 2.6. For example, Network slicing
is one of the main features of 5G networks used to optimize the allocation
of resources and increase cost and energy efficiencies. Network slicing is the
separation of multiple virtual networks that operate on the same physical
hardware for different applications, services or purposes. This form of vir-
tual network architecture combines principles behind SDN and NFV on a
fixed network to increase flexibility. In 5G networks, one physical network
will be virtually separated into multiple radio access networks and a single
network might connect several services. This model allows 5G network op-
erators to choose the characteristics needed for different capabilities such as
connection density, spectrum efficiency and traffic capacity.
The cell densification is an expected technique in 5G to improve network
capacity, especially in terms of connection speeds. The next-generation 5G
HetNet is envisioned to provide seamless connectivity across multiple RATs
such as C-V2X, Wi-Fi, DSRC and WiMAX. Switching across different RATs
disrupts any ongoing communication, resulting in frequent handovers in the
environment with the mix of macro/pico cells [3]. The new 5G-V2X system
should be designed to support advanced automotive services with the strin-
gent requirements of lower latency, higher reliability, and higher throughput.
17
(a) Autotalks Pangaea [17] (b) Bosch prototype
Figure 2.7: Examples of the V2X boards
Nevertheless, these are many issues yet to be solved. The standardization of
5G-V2X was launched in June 2018, and Rel-16, which supports the 5G-New
Radio (5G-NR) capabilities was completed in 2020 and Rel-17 is expected to
be completed in June 2021. Authors in [16] compare the physical layer perfor-
mance of NR-5G and upcoming IEEE 802.11bd V2X technologies in various
scenarios in terms of reliability. Based on simulation results, it is shown that
NR-5G can be expected to outperform 802.11bd whereas 802.11bd is severely
affected by Doppler shifts. Although 5G-V2X has good prospects, in the cur-
rent and near future standards there is no clear definition of 5G HetNet and
multi-RAT integration. On the early development stage, 5G-V2X can be
complementary to C-V2X and supports more advanced services, especially
with high communication requirements since NR-5G will not implement any
backward compatibility.
2.4 V2X equipment
In an IoV environment, every existing vehicle should be able to assume the
role of the sender, receiver, and router, to disseminate information in the
network. To allow this, vehicles should be equipped with an OBU, used
to receive and transmit messages over the air between cars and infrastruc-
tures using dedicated RAT. Figure 2.7 shows an examples of OBU integrated
solution that contains DSRC communication module, modem, security mod-
ule and software stack. To obtain the relevant data from the sensors (e.g.,
speed, acceleration), board communicates with the vehicles Electronic Con-
trol Unit (ECU) through dedicated connection, such as Controller Area Net-
work (CAN), FlexRay or BroadR-Reach physicals interface.
18
Recently, leading car manufacturers announced their intentions to equip
new lines of vehicles with V2X capabilities. Since there is no regulation
specifying how the safety critical messages should be communicated, V2X
diverged into two main standards, DSRC and C-V2X, with fundamentally
different architectures. This makes difficult for OEMs to decide which stan-
dard they should choose and even more difficult to harmonize a single global
solution. Automobile OEMs could choose the underlying V2X technology
they prefer. However, it can be anticipated that new generation OBU will
support multiple wireless interfaces, providing better communication perfor-
mance. Figure 2.8 shows the recent vehicles Advanced Driver Assistance Sys-
tem (ADAS) prototype developed by Bosch Car Multimedia, as an example
of the 2nd generation V2X board, aims to resolve the issue of the two compet-
ing standards. It was produced and flashed with a fully functional image of
the operating system and complementary DSRC and C-V2X communication
stack. Consequently, economies of scale will create a cost-effective product
which can be deployed in motorcycles and vehicles everywhere.
Figure 2.8: New generation V2X equipment supports multiple RAT
2.5 Heterogeneous communication
Several RATs were considered supporting ITS services, but the performance
requirements of vehicular networking applications are difficult to meet by a
single communication technology. Heterogeneous approach based on multi-
RAT is now essential to provide the best vehicular connectivity and can offer
many potential benefits, such as high data rates, low latency, and extended
communication range. Combining different standards brings immense op-
19
portunities as well as challenges to provide V2X connectivity that could not
only enhance existing services but also supports emerging applications [18].
Currently, the heterogeneity of existing network infrastructures has caused
challenges in network management and integration. Furthermore, heteroge-
neous networking is one of the characteristics of IoV, as well as massive data
transmission and intensive task computation.
The concept of heterogeneity defines a variety and coexistence of differ-
ent networks within one communication environment, where various wireless
technologies, such as Wi-Fi, WiMAX, LTE, can work cooperatively to sup-
port V2X capabilities. A global IoV platform that supports different tech-
nologies would be suitable for all markets and have attracted considerable
interest from the research community. Combining diverse technologies in a
hybrid approach can improve the reliability of communication systems. The
study [19] conclude that heterogeneous vehicular networking with LTE for
V2I communications and DSRC for V2V communications are one of the best
solutions for supporting vehicular services. Thus, heterogeneous vehicular
communications are expected to meet several requirements of ITS services
by integrating different wireless networks such as IEEE 802.11p and C-V2X.
Several technologies must coexist in the future IoV environment creating a
hybrid V2X approach. However, this heterogeneous approach requires fur-
ther research to find better network usage schemes.
2.5.1 Handover in V2X
Vehicular communication in a HetNet implies switching between RATs. The
V2X connectivity should be able to deal with frequent handover in dense
and small cell deployment due to the high mobility of the vehicle. Verti-
cal Handover (VHO) or inter-network handover, is the process of switching
across different technologies when connection migrates between HetNet. For
example, handover between LTE base station (eND) and Wi-Fi access point.
Handovers in a homogeneous network are referred to as Horizontal Handover
(HHO) or intra-network handover, in the sense that the next point of the
attachment of a vehicle belongs to the same RAT. For example, the RSU is
switched when the vehicle drives through a series of roadside DSRC networks
(Figure 2.9). A handover can be carried out by either the vehicle or the in-
frastructure according to some defined standard. Many parameters should
be evaluated before a handoff takes place when a vehicle moves from one net-
work area to another. Furthermore, hard handover means that the vehicle
must be disconnected from the current network before it can reconnect to
another cell or network. Whereas soft handover means that the vehicle can
be connected to more than one network during the handover process.
20
Figure 2.9: Horizontal and vertical handover in vehicular HetNet
The handover process involves three steps: handover initiation, handover
decision, handover execution [10]. It is associated with several challenges
such as handover delays, packet losses and frequent handover (ping-pong ef-
fect). The paper [1] presents an overview of VHO techniques, along with
the main algorithms, protocols and tools proposed in the literature. In ad-
dition authors suggest the most appropriate VHO techniques to efficiently
communicate in vehicle heterogeneous environments considering the partic-
ular characteristics of this type of network. Authors in [20] conclude that
for better performance, VHO decision algorithms should combine various pa-
rameters and adopt multi-criteria logic for a more effective handover process.
Midya et al. [13] review different types of VHO schemes and make a detailed
comparison based on parameters like handoff delay, decision mechanisms and
technologies used. Also, some improvements are proposed over the presented
handover mechanisms.
2.5.2 Seamless handover
In V2X, continuous connectivity is a huge challenge caused by the extremely
dynamic network topology and the variable number of mobile nodes. In
VANET, connections often suffer in terms of packet loss, delays and low
throughput. To provide reliable ITS services a seamless connectivity (i.e.
continuous connectivity) in heterogeneous IoV environment is required. In
the context of multi-RAT usage, seamless vehicle communication is one of
the top priority objectives. Currently, many studies have investigated the
seamless connectivity in heterogeneous environments [1, 8]. The traditional
21
V2X heterogeneous connectivity relay on switching from one technology to
another and may disrupt any ongoing connection. Therefore, a handover
mechanism is required to guarantee QoS with continuous connectivity. Sup-
porting seamless connectivity for different V2X services presents several chal-
lenges and can only be assured using the advanced techniques discussed in
Chapter 4.
22
Chapter 3
Challenges and opportunities
To better assist future IoV, V2X communication should deal with issues such
as dynamic topological changes, random network density, interrupted connec-
tion, spectrum scarcity, and computation resource limitation [7]. Delivering
high speed and reliable V2X connectivity by integrating multiple RATs is a
challenging task, since these technologies have been developed for different
types of networks and their characteristics are different. The interoperability
of radio access technologies remains an important area of research. There-
fore, based on the thesis objectives presented on Chapter 1, the first set of
research questions can be defined in the scope of this PhD work. Can multi-
RAT capabilities improve vehicle connectivity? That is, can the integration
of different wireless technologies potentially improve V2X in terms of relia-
bility, throughput, delay, etc., and to provide applications with the required
level of QoS? Is hybrid approach can deliver more data reliably, with lower
latency while avoiding unnecessary switching between RATs? How to assign
a particular vehicular user to the most suitable network and how the link per-
formance will change when the connection migrates between networks? The
major challenge that remains an open research problem for multi-RAT usage
is seamless connectivity and automation aspects. If the effectiveness of such
heterogeneous approach for vehicle communication will be confirmed, the fol-
lowing question arises: how to implement seamless V2X in a heterogeneous
environment? This question can be subdivided into:
A. How to implement a handover decision algorithm, i.e. how to make
a selection of a suitable RAT? The decision algorithm address to two
important questions:
(i) Whether to perform handover or not? Or, how to select suitable
time to trigger the handover?
23
(ii) To handoff to which network when there are various options avail-
able? What a network selection algorithm should be applied that
guaranties better connection performance and improves the over-
all throughput of a network? What proper metrics and methods
should be adopted for the selection algorithm?
B. How to implement a handover procedure? That is how to switch from
one radio access technology to another induced by the decision algo-
rithm. How to preserve connectivity and QoS during the handover
process? How to reduce the impact of the signalling overhead when a
handover is performed?
The development of V2X heterogeneous connectivity has considerable
challenges such as ensured fairness, seamless connection, QoS requirements
support, resource allocation, interference, coexistence and integration of dif-
ferent wireless technologies. Consequently, to support IoV requirements,
the solution should be designed using recent mechanisms, devising a control
paradigm that can minimize the impact of frequent handovers, provide the
QoS, reliability, availability, and security. The solution should be transpar-
ent to applications and it also has to ensure that new standards and services
can be integrated easily. In the next section, we select the most promising
approaches and future research directions for heterogeneous V2X.
24
Chapter 4
State of the art
Selecting the most suitable available network technology is important to
maintain a satisfactory QoS for any ongoing communication. Optimal selec-
tion of underlying networks in a heterogeneous V2X environment is always
challenging and depends on various parameters. Therefore, the selection of
proper metrics and methods for the handover decision in a vehicular net-
work is an important area of research. The VHO process can be initiated
according to user choice or application requirement. The handover process
in V2X relies on several parameters, such as Received Signal Strength In-
dicator (RSSI), movement direction and speed, network bandwidth, traffic
load, delay, usage cost. Furthermore, the handover decision can be influenced
by upper layer applications since V2X services have diverse requirements in
terms of latency, bandwidth, delay variation, and error rate [10].
This chapter presents the state of the art solutions that leverage inte-
gration of multiple RAT s and briefly describe VHO techniques for HetNet.
It analyzes the research effort in the field of hybrid communication and ap-
proaches presented by the research community and industry. The adopted
methodology in this study was the literature research on solutions for HetNet.
Some keywords especially used to search on the proposed context for this
thesis were V2X, C-V2X, DSRC, VANET, heterogeneous networks, IEEE
802.11p, RAT selection, (VHO), IoV, vehicular networks, 5G, and others.
The most consulted digital libraries were IEEE, Elsevier and Springer with
around the 80% of selected references matched to the period from 2018 to
2020. In this literature research more than 150 studies were analyzed and
the most relevant were classified into seven main categories based on their
implementation of hybrid connectivity mechanisms. Furthermore, the review
presented in this chapter discusses the main open challenges and opportuni-
ties and offers a good basis for next research and development activities.
25
4.1 Review of V2X domain
Several excellent surveys on the vehicular network are presented in litera-
ture. The work in [8] is a new contribution in a similar category of tutorials
presenting latest details of the V2X domain. It is a comprehensive survey
of V2X that covers the state of the art and future research directions. Also,
authors describe the architecture, applications, emerging RATs, standard-
ization, and project activities in ITS. Challenges associated with multi-RAT
usage to fulfill the unique V2X requirements are covered as well. Studies in
[21] review the state of the art of IoV applications in smart city, presenting
a comprehensive study of scientific and industrial progress in this field, in-
cluding a representative collection of the most cited papers published in the
literature.
The survey in [1] presents the basic characteristics of V2X communication
in HetNet environments. The research provides a comprehensive overview of
the V2X including different applications, research initiatives, characteristics
and requirements for vehicle communications in a heterogeneous environ-
ment. Also, authors identify several V2I key research challenges; surveys
and reviews some related studies in the literature that explores VANET het-
erogeneous communications in terms of vertical handover, data dissemination
and collection, gateway selection and other issues. At the end a list of rec-
ommendations for the possible R&D activities is summarized.
The work in [10] presents a comprehensive study on heterogeneous V2X.
Furthermore, a comparison of various solutions proposed by the research com-
munity in terms of handover, data dissemination, gateway selection, QoS and
other concerns was presented. A recent study in [7] provides a comprehen-
sive survey on V2X technologies; highlighting the key technical challenges
and opportunities toward IoV, as well as demonstration of emerging R&D
directions. Another new survey [3] presents a complete research on 5G tech-
nology evolution and infrastructure associated with V2X communication by
IoV, considering its architecture, applications and the V2X features and pro-
tocols. Studies in [22] examine the most relevant V2X systems, applications,
and communication protocols that will distinguish the future IoV environ-
ment used by vehicles. Different types of V2X communication are analysed
separately to highlight in more detail the distinctiveness of each of them.
Paper in [6] presents a comprehensive framework of IoV with emphasis on
layered architecture, protocol stack, network model and challenges. Follow-
ing the background on the evolution of VANETs and motivation on IoV, the
author’s overview of IoV is presented as a vehicular HetNet. A five layered
architecture of IoV is proposed considering functionalities and representa-
tions of each layer. A network model of IoV is proposed based on the three
26
network elements, including cloud, connection and client. The benefits of the
design and development of IoV are highlighted.
Several works have recently discussed the two main V2X technologies
(i.e., C-V2X and DSRC) from various perspectives and numbers of perfor-
mance comparisons have been provided. Authors in [11] aim at revising
the performance comparison between IEEE 802.11p and C-V2X Side link,
first summarizing the related work and then providing evaluation in various
realistic scenarios. Based on original results obtained through large-scale
simulations, C-V2X expected to provide a longer range than the standard
IEEE 802.11p and overall significant performance improvement in most sce-
narios. However, note that tests with real hardware are still at an initial
stage. Study in [4], an overview of the two main V2X standards, DSRC and
C-V2X, their core parameters, shortcomings, and limitations, and explore
the need for integration of IoV-based solutions. Overview and specifications
of commercially available products is also provided.
4.2 Standard Approaches
As mentioned before, the requirements of IoV pose a big technical challenge;
therefore, exploiting one RAT is not sufficient to support V2X services. To
support seamless vertical handover two mechanisms were proposed indepen-
dently by IEEE and 3GPP, namely; Media Independent Handover (MIH)
and Access Network Discovery and Selection Function (ANDSF).
To cope with the challenges of the HetNet, IEEE 802.21 proposed in 2009
the MIH standard to enable the handover of IP sessions from one RAT to
another such as Wi-Fi and cellular technologies. The MIH is a sublayer be-
tween Layer 2 (L2) and Layer 3 (L3), which provides an optimized handover
mechanism, selecting the best network based upon the information gathered.
The middleware protocol called media-independent handover (MIH) func-
tion allows the handover management process to operate independently of
the data link layer and physical layer, providing the services to upper layers.
MIH facilitates network discovery and selection procedures, providing the list
of available networks with the relevant information that allow the vehicle to
select more suitable RAT. Thus, seamless handover for any moving device in
HetNet can be achieved. However, in the context of V2X the list of candidate
RATs should represent networks toward vehicles moving direction, otherwise
it will spend much time in scanning all available candidates and will cause
high handover latency. The MIH standard does not require additional en-
tities to provide a handover; hence, most VHO approaches founded in the
literature are based on this mechanism. However, MIH requires a modifica-
27
tion on legacy radio systems since there is a need for measurement reports
between RATs that advise the best network. The work in [23] concludes that
the current MIH standard needs modifications to fit IoV communication re-
quirements in terms of latency and packet loss.
ANDSF mechanism facilitates the seamless VHO between the 3GPP net-
work and other types of wireless technologies (e.g., Wi-Fi) that allow a mobile
node to offload traffic directly to the non-3GPP access network. It assists to
automatically discover, select and connect the most suitable wireless network
based on certain policies that are predetermined by network operators. Cellu-
lar core network provides to the mobile node, a list of radio access networks
that is available in its vicinity, along with their deployment coordinates.
It can help vehicles preselect the target network for handover procedure.
Further enhancements of the standard provide a mechanism for intelligent
network selection and traffic steering with multiple IP connectivity through
trusted WLAN. However, ANDSF needs additional entities (e.g. ANDSF
Server) to provide handover between RATs and mobile devices can be at-
tached to either cellular or Wi-Fi, not to both at any given time. Moreover,
the current solution has a high signalling overhead and incurs high latency,
thus further work is needed to develop a mechanism to support high-speed
mobile nodes in multi-tier heterogeneous IoV environments.
The implementation of MIH and ANDSF protocols are still in progress
and the detailed description of handover mechanisms of both can be found
in [23]. Hence, reliable seamless handovers across different RATs are still
unavailable today.
4.3 Network Selection
A vehicle on the move should find the most suitable network in real-time,
therefore it is essential to consider multiple parameters when selection is
made. Many different VHO methods can be found in the literature. Among
other solutions, the Multi-Criteria Decision Making (MCDM) model is fre-
quently used due to its simplicity and diversity of network selection algo-
rithms. MCDM algorithms shall combine multiple parameters to make net-
work selection and stable vertical handover. MCDM techniques rank avail-
able networks based on multiple parameters, like network traffic load, vehicle
speed, type of service, RSSI, network cost, etc., and then select the most ap-
propriate underlying wireless network to connect to [1]. An overview of VHO
techniques were presented in [24], along with the main algorithms, protocols
and tools proposed in the literature. Authors classified the most widely used
VHO decision algorithms in the literature into different sets of algorithms de-
28
pending primarily on the information used to make decisions. Furthermore,
authors consider the importance of a fast MCDM algorithm to perform an
accurate decision, and to select the best candidate network.
Mir et al. [25] present hybrid approaches suitable for heterogeneous V2X
where a QoS-aware RAT selection algorithm is proposed and evaluated. The
algorithm selects between IEEE 802.11p and LTE network by considering
network load and application’s requirements. The results based on simu-
lation studies show that the proposed RAT selection mechanism results in
significant communication performance improvements. In a recent study [18],
the same authors propose a protocol stack for hybrid vehicular network archi-
tecture, which combines DSRC and C-V2X technologies. A distributed radio
resource management entity was designed and evaluated. It employs conges-
tion control based on locally available measurements. A generic framework
called CellCar is presented for dynamic RAT selection and communication
management in hybrid vehicular networks. The simulation results show the
effectiveness of the proposed architecture and protocol suite under various
parameter settings and performance metrics such as the number of VHOs,
packet delivery ratio, throughput, and latency.
Authors in [26] present performance guaranteed optimized handover de-
cision algorithm. The process of decision making, the data rate of handover
vehicles is estimated and simulations are performed to demonstrate the ef-
ficiency of the algorithm. Jacob et al. [27] highlight that none of the tech-
nologies is flexible and reliable enough to serve diverse requirements in terms
of delay, reliability and throughput for vehicle communications. This work
presents the latest overview of the potential, challenges and main design as-
pects of hybrid V2X communications. It also describes newly introduced
Hybrid Communications Management (HCM) - a control function that se-
lects the access interface based on communication profile of the application.
It is an independent component bridging between the access interfaces and
network stack (Figure 4.1). The HCM enables Multi-RAT coordination per
packet-level and allows for easy implementation and integration into hybrid
connectivity modules based on standard hardware and software components.
The same authors in [28] present a Multi-RAT redundancy protocol, the core
element of HCM that increases the reception probability by sending multiple
copies of the same packet over different RATs.
Context-Aware Heterogeneous V2X Communications architecture was
presented in [29]. The proposed solution allows dynamic selection and config-
uration of communication profile based on context conditions (environment
conditions) and the application requirements. The study considers server-
based architecture and the vehicle nodes are in charge in requesting the
context information they need. The implementation and evaluation of a
29
Figure 4.1: Radio interface selection function
heterogeneous V2I communications algorithm was performed and the poten-
tial of the proposed architecture is demonstrated. The same authors in [30]
propose the first decentralized context-aware heterogeneous V2V communi-
cation algorithm that is technology and application agnostic and allows each
vehicle to autonomously and dynamically select its communication technol-
ogy, considering its application requirements and the communication context
conditions. This study considers a multi-link and multi-RAT vehicular sce-
nario where all vehicles are equipped with different RAT’s interfaces. The
proposed algorithm works locally at each vehicle to seek an optimal solution,
considering the decisions previously taken by its neighboring vehicles, and
the impact that its decision could have on its neighbor vehicles.
Authors in [31] propose a VHO algorithm for V2I communication. The
proposed solution is empowered by the MIH standard, which selects the
most suitable technology considering the particularities of vehicular networks
such as, surrounding context, application requirements and user preferences.
Therefore, a MCDM algorithm selects the network that best meets the end-
user connectivity requirements. Paper in [5] evaluates V2X system perfor-
mance considering C-V2X exclusively. Authors propose a communication
scheme where the data pass through both the LTE-Uu and PC5 interfaces
to obtain a diversity gain and improve system reliability. The proposed so-
30
lution was evaluated in different scenarios using a simulation environment.
The recent study in [32] presents a hybrid MCDM algorithm. The solution is
based on the so-called hybrid algorithm combining several simpler handover
decision mechanisms, available in literature, to bring together their advan-
tages by the method of linear combination. Authors focus on applying a
hybrid algorithm to the complex problem of optimal network selection with-
out considering the network load balancing. Table 4.1 summarizes described
approaches for network selection.
Author Year Approach Networks
Mir et al.[25] 2015 QoS-based DSRC, LTE
Mir et al.[18] 2020 Distributed,
QoS-Based
DSRC, C-V2X
Jacob et al.[27] 2018 Multi-profile,
dynamic
DSRC, LTE-V2X
Jacob et al.[28] 2019 Multiple copies DSRC, LTE-D2D
Sepulcre et al.[29] 2018 Centralized
Context-Aware
DSR, LTE, WiFi
Sepulcre et al.[30] 2019 Distributed
Context-aware
DSRC, C-V2X
Yu et al [32] 2019 Hybrid, multi
attribute
3G, LTE, WiFi
Barja et al. [31] 2015 QoS-Based,
Context-aware
Wi-Fi, WiMAX, UMTS
Lianghai et.al.[5] 2018 Multiple copies LTE, LTE-D2D
Table 4.1: Proposed network selection algorithms
4.4 IP-based solutions
The common part of all existing and future heterogeneous wireless networks
is the Internet Protocol (IP) protocol. Thus, the IP layer could be the best
location to deploy QoS mechanisms for unified multi-RATs technology. We
have identified many different existing approaches based on Internet mobil-
ity management protocols Mobile IPv6 (MIPv6) to provide V2I communi-
cation [1]. Among other existing solutions, Proxy Mobile IPv6 (PMIPv6)
is a promising protocol in this field. The network-based mobility protocol,
PMIPv6, is the enhancement of MIPv6 that enables the mobile node to
31
change its network without any signalling being generated. PMIPv6 has a
centralized architecture, i.e., all the mobility control signals and data traffic
pass through a centralized anchor node.
To take advantage of all radio interfaces of the vehicle, work in [33] present
seamless flow mobility management architecture based on network applica-
tion classes and mobility management. The solution deals with several net-
work interfaces at the same time trying to maximize the network throughput,
and to satisfy minimum requirements of latency and packet loss for each class
of vehicular network application. Simulations were performed and the results
showed a low handover time, delay and packet loss values. Additionally, a
new hybrid interworking scheme was proposed in [34], which enables access
to mobile internet and general IP services over a mobility management mech-
anism. Furthermore, the solution focuses on urban vehicular scenarios and
enables seamless communications regardless of roaming agreements between
network operators. This novel hybrid scheme allows a seamless transfer of
IP sessions, despite different patterns of mobility and the heterogeneity of
the supporting radio access technologies. Performance analysis has shown
that the offered solution outperforms other protocols such as the optimized
version of MIPv6.
Dias et al. [35] proposed a mobility approach that integrates extended
mobility protocols based on PMIPv6, with a mobility manager that provides
seamless communication between vehicles and infrastructure. This method
can select the best technology to maintain the vehicle connected without
breaking any active sessions. The proposed architecture deals with both
Layer 2 and Layer 3 handovers. For L2 handover, a mobility manager that
scans the available networks and triggers the handover was designed; whereas,
for L3 handover control, the mobility protocols were enhanced and modified
to be coupled with the mobility manager. Experiments were done in real
vehicular environments combining three technologies: IEEE 802.11p, IEEE
802.11 g, and 3G. The results demonstrate the advantages of a simplified
communication standard for VANETs since the traditional Wi-Fi standards
introduce high handover latency and packet loss, being it critical to deploy
IEEE 802.11p-enabled RSUs in high demanding scenarios. The results also
show that if IEEE 802.11p is used in both vehicles and RSUs, the proposed
approach can perform seamless handover with low delay and no packet loss.
Enhancement of Fast Handover Mobile IPv6 (FMIPv6) [36] was presented
and is based on a handover management technique using the concept of
tunneling in a VANET scenario. Numerous parameters such as handover
latency, signaling overhead, performance comparison using tunneling, packet
loss, service disruption time and network lifetime were taken into considera-
tion. Hence, there is a need to evaluate these techniques in a more realistic
32
Author Year Highlights
Meneguette
et al. [33]
2013 Trying to maximize QoS by dealing with several
network interfaces. Flow mobility management ar-
chitecture proposed
C´espedes et
al. [34]
2015 Investigate the seamless Internet access over the
HetNet. Propose a hybrid global mobility scheme
that allows for IP sessions to be transferred across
dissimilar radio access networks
Dias et al.
[35]
2012 Method provides seamless V2I communication.
Mobility manager that scans the available net-
works and triggers the handover was designed
Dahiya et al.
[36]
2014 Evaluated the use of enhanced FMIPv6 for ve-
hicle communication. Numerous parameters are
considered
Tuyisenge,
L. et al. [37]
2020 Allows a vehicle to simultaneously connect to any
available network. Method provide seamless VHO
using logical interface and link-layer multiplexing
Murtadha,et
al [38]
2016 Fully decentralized approach based on the cross
layer design
Table 4.2: Network-based vertical handover approaches
scenario and applying them to an actual wireless situation.
Some primary drawbacks exist in IP-based approach. For instance, PMIPv6
presents a large overhead on the time and wireless resources within the ra-
dio access network. It is facing the disadvantages of centralized systems
such as overhead, bottleneck, single point of failure and no scalability. Fur-
thermore, PMIPv6 might present high handover latency depending on the
distance between the vehicle and core networks, and disconnection periods
due to exchanging the time of handover messages and new address configu-
ration time. The research community proposes many different improvements
in PMIPv6 protocol, for example Fully Distributed Mobility Management
scheme [38]. The proposed approach removes any central entity in the net-
work infrastructure by moving the mobility functionalities closer to the user
and distributing the control and data planes at the edge of the access net-
work. Moreover, the solution uses MIH framework to provide seamless L2
handover in heterogeneous wireless networks.
Recent work in [37] proposes a VHO mechanism, denoted Proxy MIPv6-
based Mobile Internal Vertical Handover (PMIP-MIVH), which uses a logical
33
interface and a Distributed PMIPv6 scheme to improve the handover per-
formance. The most important advantage of a logical interface is that it
may be bound to multiple physical interfaces by a so-called “link-layer mul-
tiplexing” in order to provide alternative communication paths. The pro-
posed mechanism was evaluated using a network simulator and the results
show that the presented solution outperforms existing PMIP-based solutions.
Proposed method allows the vehicle to directly and simultaneously connect
to any available network that ensure the session continuity (seamless VHO)
and consequently reduce the handover latency and packet loss. Table 4.2
summarizes described solutions founded in the literature highlighting their
main futures.
4.5 Multipath Solutions
Most of the end-user devices are now equipped with multiple network inter-
faces such as cellular and Wi-Fi. IoV applications require continuous and
seamless connectivity, which is a big challenge in the context of vehicular
environment and multiple available RATs. The V2I normally uses only one
interface for connectivity at any instant of time and relies on the handover
mechanisms when it is necessary. Traditional transport protocols cannot
benefit from multi-RAT availability, since data are forwarded through a sin-
gle path, even though there are many available paths between the source and
destination. Furthermore, in the context of V2X the traditional data deliv-
ery paradigm suffers from high delays and packet losses caused by frequent
handover in vehicular environment.
Figure 4.2: MPTCP position in protocol stack
34
Now, Multipath TCP (MPTCP) [39] technology is gaining momentum to
meet the requirements of TCP-based services in HetNet. It is a new trans-
port layer protocol that enables the use of multiple network interfaces and
IP paths simultaneously with the ability to aggregate the available band-
width of several links given more network resources to flow by splitting them
into multiple paths. These multipath solutions can replace VHO technolo-
gies allowing vehicles to connect to different available RATs simultaneously.
MPTCP can provide seamless connectivity across vehicles HetNet and sig-
nificantly reduce VHO and delay associated with it. It provides resilience to
failures i.e., when one link goes down transmits data on alternate paths.
The fact that MPTCP can use all available network interfaces simultane-
ously may improve vehicles connection performance. MPTCP is backward
compatible with TCP and provides at least the same as what is best possi-
ble when the connection is a single path. MPTCP layer is located between
the application and network layers (Figure 4.2). It aggregates several TCP
connections using sub flows that are similar to the normal TCP connection
with some additions. However, the most significant advantage is that there
is no need to make changes in the legacy backbone infrastructure; the only
requirement is that the end devices should be MPTCP capable, Figure 4.3.
Figure 4.3: MPTCP in vehicular HetNet communication scenario
Multi-path solutions are not much explored in the IoV context. A study
in [40] investigates the performance of MPTCP for V2X under distinct ve-
locities in LTE and Wi-Fi access networks. The real test was performed to
evaluate the multipath solution. The results show that MPTCP allows robust
35
connections with seamless handovers while it maintains comparable perfor-
mance to TCP. However, MPTCP does not currently operate well under very
high vehicle velocities in V2V configuration. Williams et al.[39] study the
performance of MPTCP for vehicle communications. Experiments were per-
formed on a physical testbed and small-scale vehicle-based field, simulating
MPTCP mobility using WiFi and 3G. Authors found that using MPTCP
across paths of similar characteristics, download times are in most cases at
least as good as single-path TCP over the best link. Nevertheless, a result
shows that MPTCP can perform worse when paths are asymmetric (in terms
of RTT, and bandwidth).
Work in [41] is the first study in which both MPTCP and Software De-
fined Networking (SDN) are explored for V2I communication across small
cell (Wi-Fi and DSRC) deployments of a smart city. Authors investigate
the impact of L2 handover and explore the role of SDN for improving the
overall performance of MPTCP in a vehicles HetNet scenario. First, au-
thors evaluate the MPTCP performance in SDN architecture with vehicular
mobility scenario using a simulation tool in terms of packet loss, RTT, and
throughput. Then, authors propose an enhanced rule installation mecha-
nism that can predict the next point of association of vehicles and trigger
SDN controller to install rules (in data plane) in advance. Authors expect
that proactive flow creation can reduce the load on the controller and help
minimize congestion on the common physical control channel. The study in
[42] investigates MPTCP algorithms for reliable communication over vehicles
HetNet and proposes MPTCP-IoV algorithm. Authors perform load balanc-
ing and forward error correction (FEC) techniques to use path diversity and
enable reliable communication over multiple parallel HetNet, while satisfy-
ing delay constraints. In addition, a comprehensive mathematical analysis is
conducted.
Multipath IP (MPIP) [43] was introduced in 2017, and unlike MPTCP,
MPIP is also compatible with UDP-based applications. It controls multi-
ple paths in the network layer that require additional mechanisms to feed-
back end-to-end path information and a multipath IP-routing method. Re-
searchers have claimed that MPIP can efficiently use all available interfaces
and applications can benefit from multipath transmissions in terms of higher
aggregate throughput. It is resilience to failures and traffic variations on indi-
vidual paths and can seamlessly work across different networks. There is no
need to change the upper-layer protocols. MPIP can adjust the transmission
strategies to satisfy diverse application needs. It can customize its multipath
routing depending on the application and user needs and like MPTCP, it
only requires changes at end devices.
Paper in [44] discusses the performance of MPTCP and MPIP since these
36
methods cannot improve the connectivity because scheduling to multipath for
concurrent applications is not effective, i.e., flows off all applications to each
interface. Authors propose the path-selection method based on the available
bandwidth on interfaces by controlling the paths in each application of end
devices. This method monitors the number of flows through each interface
of a device, and calculates the estimated bandwidth for each interface. The
proposed method was evaluated using a simulation tool. Results show that
the solution can effectively use resources and improve the performance of all
applications.
Quick UDP Internet Connection (QUIC) is a recent connection-oriented
protocol proposed by Google that combines the functions of HTTP/2, TLS,
and TCP protocols into a single application layer protocol that runs over
UDP. Work in [45] combines the multi-streaming capability of QUIC proto-
col with SDN to forward the data through multiple paths available in the
network. However, this solution has a distinct feature of allowing the source
and destination with single network interface cards.
Motivated by the success of MPTCP, authors in [46] design Multipath
QUIC Protocol (MPQUIC). It allows to create a connection using a different
RAT simultaneously; thus, allows dual-homed hosts to send QUIC traffic
through multiple paths. MPQUIC has the same benefits as MPTCP, such
as multi-interface aggregation, and can perform network handover by design.
However, authors show that MQUIC can provide improved benefits to QUIC
than Multipath TCP to regular TCP and copes better with packet losses.
MPTCP and MPIP can replace vertical handover technologies, allowing
vehicles to use different RATs simultaneously. However, the probability of
packet loss can be high and packets arrive at the destination out-of-order
due to time-varying heterogeneous wireless paths. Further, the IoV traffic
is delay sensitive, which urges the need to investigate multipath algorithms
for reliable communication over HetNet, while satisfying delay constraints.
Multipath approaches have the potential to automate driver‘s experience
through improved resilience to wireless network failure as well as improved
throughput. However, these technologies are not much explored in the IoV
context, thus, there is an excellent opportunity for its use and improvement
of V2X communication. Various cases for vehicle multipath communication
are summarized in Table 4.3.
4.6 Software-defined networking
Software Defined Networking (SDN) is a network paradigm designed to solve
problems of traditional networks by introducing a programmable mechanism
37
Author Year Protocol Highlights
Mena et
al.[40]
2017 MPTCP Describes a testbed that studies the
performance of MPTCP under VANET
configurations using in LTE and Wi-Fi
networks
Pokhrel et
al. [42]
2019 MPTCP Approach challenges of MPTCP usage in
IoV. Load balancing with forward error
correction mechanism was presented for
performing coupled congestion control in-
side MPTCP
Williams et
al [39]
2014 MPTCP Characterise the V2I performance of
MPTCP using a mix of wireless links.
Field test performed using 3G and a
DSRC
Sun et al.
[43]
2018 MPIP Authors present a MPIP protocol as alter-
native for MPTCP on the network layer.
Design and performance evaluation also
presented
De Coninck,
Bonaven-
ture [46]
2017 MPQUIC Present and evaluate a MP QUIC pro-
tocol by comparing them with TCP and
MPTCP in a variety of settings
Ishida et al.
[44]
2018 — Propose the path-selection method that
assigns a flow of each application to a
certain interface based on the available
bandwidth
Rezende et
al. [45]
2019 QUIC +
SDN
Design and implementation of an SDN-
based framework for routing multi-stream
traffic over multiple paths for single-
homed devices
Singh et al.
[41]
2019 MPTCP
+ SDN
Explore the role of SDN in improving
the overall performance of MPTCP in a
vehicle HetNet scenario. Proposes en-
hanced rule installation mechanism for
SDN controller
Table 4.3: Multipath approaches for vehicular HetNet
to manage and configure every device on the network with a central SDN
controller. It provides the global view of the entire network and can control
38
network traffic flexibly by separating the control and the data planes. The
control plane and data plane can communicate with each other with a unified
interface (i.e., API). Therefore, it simplifies network management, and offers
programmable and flexible network architecture for IoV. A network model
based on the SDN computing paradigms has higher flexibility, scalability,
location capability, and fewer delays than the current network models. SDN
networks allow heterogeneous communication through their characteristics
of reconfigurable and reprogrammable networks. Furthermore, devices from
various vendors can communicate with each other via a standardized inter-
face. SDN is a promising candidate for implementing an IoV, which enables
the coexistence of various RATs. A topological architecture of vehicular
HetNet based on the SDN system is depicted in Figure 4.4.
Figure 4.4: A topological architecture of software-defined vehicular HetNet
We found several studies in literature, which have explored SDN for ve-
hicular networks. The work in [47] presents a comprehensive survey and
recent advancements of existing vehicular SDN systems. Work focuses on
the use of SDN paradigm in VANET along with opportunities and chal-
lenges of support for vehicles HetNet. Authors compare the implemented
protocols for software-defined vehicular communication presented in the lit-
erature, covering different scenarios and simulation tools. A recent paper in
[48] addresses the challenges related to SDN-based architecture to serve the
services of ITS using a vehicular network. Authors investigate state-of-the-
art SDN architectures in vehicular networks describing the SDN paradigm
and the requirements and discuss some challenges in these architectures.
39
The study in [49] propose cloud-based architecture for soft-defined vehic-
ular HetNet. Authors discuss the opportunities and challenges of applying
SDN approach to a vehicular network. They investigated its feasibility in
the emerging 5G era and proposed a new hierarchical control layer to enable
a unified and flexible network. The communication resources are managed
by communication control functions that make the coordination of physi-
cal and MAC operations for heterogeneous network devices. The network
architecture presented can improve network efficiency while reducing the
costs of management and maintenance. Alternatively, He et al. [50] ad-
dressed the data transmission problem in vehicular HetNets and proposed
a SDN-based heterogeneous communication coordination approach. With
SDN-based wireless communication solutions the network resources can be
well managed to reduce communication cost.
Paper in [51] reviews various approaches for multi-RAT control and pro-
poses novel SDN-based scalable network architecture for unified control and
management of diverse technologies. The architecture provides end-to-end
network control while ensuring scalability through the creation of multiple
logical networks (or network slices) over a single physical network. The result-
ing architecture provides a framework for the deployment of applications in
a RAT agnostic fashion by abstracts RAT-specific details. The performance
was evaluated in simulation showing improved overall network performance.
Work in [50] describes SDN-based architecture for vehicular communication
in HetNet, where all network nodes are abstracted as programmable SDN
switches and configuration of wireless devices and network resources is per-
formed by a centralized control plane. Presented architecture is divided into
three layers, i.e., the data plane, control plane, and application plane. More-
over, to mitigate the SDN management overhead, a trajectory-prediction-
based vehicle status update policy was designed. The effectiveness of the
proposed architecture was validated through a simulation.
Authors in [52] confirm that when each RAT is controlled by a set of dif-
ferent entities, leads to inefficient network resource usage, and proposes SDN
vehicle HetNet architecture for ensuring a highly agile networking infras-
tructure. In the presented solution, centralized intelligence in a single SDN
controller is augmented with localized intelligence to avoid a single point of
network failure. The abstraction of physical radio resources of diverse RAT
in terms of their bandwidth, time and location is suggested. Since the SDN
controller has global knowledge of the available resources of each base station
or AP along the vehicle’s traveling direction, intelligent VHO schemes can be
accordingly employed for switching purposes. Furthermore, authors present
a brief overview of ITS along with a debate on the potential challenges in
the implementation of the V2X connectivity.
40
A handover is always accompanied by connection changes, such as a
change of IP address, which makes it difficult to have a stable wireless con-
nection, especially at the transport layer. Paper in [53] proposes a seamless
handover scheme where SDN is used to adapt the frequent network change in
VANETs. SDN controller provides a global view of the network and conducts
a handover by maintaining the same IP address. When a handover occurs,
the data is cached on the server at the edge of the mobile network, that is,
near to the mobile node before a handover happens. This caching mecha-
nism reduces the packet loss during the handover process, so the vehicle can
restore normal communication faster.
SDN may have a significant impact on the design and development of
wireless multi-RAT networks. For unified control and management of di-
verse RATs an intelligent network architecture needs to be adopted. Such
intelligence is possible only if you have a bird’s eye view of the entire vehicu-
lar network architecture. SDN takes the network intelligence to a centralized
SDN controller. Since the SDN controller can have global knowledge of the
available physical radio resources of each network along with vehicle’s mo-
bility information such as speed, location, an intelligent multi-RAT control
can be accordingly employed. However, this domain is new and wide open
for research. Table 4.4 summarizes the research efforts on SDN solutions for
vehicle HetNet.
4.7 Cognitive radio
In the future, the ITS frequency bands (5.8 to 5.9 GHz) could be shared with
other applications, which can lead to crowding of the spectrum. Cognitive
Radio (CR) enables a more efficient use of the spectrum bands thus, improves
vehicular communication efficiency. CR was defined as an intelligent wireless
communication system capable of being aware of its environment, learning,
and adaptively changing its operating parameters in real time for providing
reliable communication and efficient use of the radio spectrum [54]. This ap-
proach can autonomously make decisions using gathered information about
the radio frequency environment. It can enhance the performance of exist-
ing V2X communication by integrating artificial intelligence with Software-
defined Radio approach (SDR).
SDR is the primary enabling technology to support implementation of
CR. It allows flexible radio operations by using reconfigurable software imple-
mentation of conventional radio components (Figure 4.5). Wireless systems
implemented by software routines allow to support various RATs by the same
hardware based on software changes. It enables vehicles to fit their trans-
41
Author Year Highlights
Meneguette
et al. [48]
2020 Recent state-of-the-art of SDN in vehicular net-
works. SDN paradigm, requirements and chal-
lenges are discussed.
Duo et al.
[53]
2020 Proposes a seamless handover scheme where SDN
are used to provide a global view of the network.
The data is cached on the server at the edge of the
mobile network when a handover takes place.
Manjeshwar
et al. [51]
2019 SDN-based framework for multi-stream transport
protocols in multipath networks. Consists by a
slice manager which splits physical networks into
multiple logical networks. Deployment of RAT ag-
nostic control applications.
Bhatia et al.
[47]
2019 Survey of recent vehicular SDN systems. Focuses
on the use of SDN paradigm in VANET along with
opportunities and challenges of support for vehicle
HetNet.
Zheng et al.
[49]
2016 Preliminary study on soft-defined vehicular Het-
Net, architecture and challenges. A new hierarchi-
cal control layer is proposed.
Z. He et al.
[50]
2016 Describes SDN-based architecture for vehicular
HetNet where all network nodes are abstracted
as programmable SDN switches. A trajectory-
prediction-based vehicle status update policy was
designed and evaluated in simulation.
Mahmood
et al.[52]
2019 Proposed architecture offers abstraction of net-
work entities as SDN switches thus facilitates
an efficient orchestration of the network re-
sources. Discusses the architectural design and
open challenges.
Table 4.4: An SDN approaches for multi-RAT usage
missions and adaptively select the communication technology to deal with
the fast changes in the radio environment. Furthermore, SDR will prevent
the need to implement hardware upgrades with emergence of new protocols
in the future. SDR-enabled hardware can switch across different RATs de-
pending on its requirements and context. CR can explore underused spectral
resources and opportunistically use the available licensed spectrum as long
42
Figure 4.5: Software-defined radio approach
as primary users are not perturbed. For instance, vehicles can use licensed
TV bands according to the requirements of the applications.
A study in [54] provides recent advances and open research directions on
using cognitive radio technology in VANETs. The motivation leading to CR
vehicular network and SDR-based communication architecture is provided.
A taxonomy of recent advances in CR for vehicular networks, including ma-
chine learning and spectrum management approaches is presented. Questions
related to the routing, mobility management and content distribution in CR-
based VANETs are also discussed. Authors in [55] survey novel approaches
and current research challenges associated with the use of CR technologies
in VANETs. It describes the benefits and potential of CR technology in ve-
hicular networks. The open issues and current research directions for future
development of CR-based vehicular networks are presented. Security and
privacy issues for CR enabled VANET are also examined.
Haziza et al. [56] propose multi-technology cooperative ADAS based
on integration of SDR devices for seamless V2X connectivity across IEEE
802.11p and LTE. The key innovation of the project lies in the design of the
multi-technology On-Board Unit (OBU) that coordinate usage of different
protocol stacks on SDR device. The performance study based on a simulation
shows that a designed solution can balance the generated traffic over two
wireless networks. It can support both LTE and IEEE 802.11p signals on the
same board and both protocol stacks are capable of working simultaneously
using different RF chains.
43
Chapter 5
Work Plan
This section describes the expected future work, proposing a work plan in
order to achieve the thesis goals. This PhD work starts with a one year
curricular plan within AESI’s doctoral program. This phase will also include
the initial research needed for the writing of the state of the art. A survey of
existing solutions of heterogeneous vehicle communication should be done on
this phase. The main objectives of this PhD work should be defined as well
(Chapter 1.1). It will end with the pre-thesis presentation and defence. The
available HetNet standards shall be identified and studied. The solutions
found in the literature should be evaluated in order to find the best suitable
approaches for vehicular HetNet scenario. In this phase a solid background
should be created for future extended research. A scientific paper should
be published presenting the work done, resulting from the literature review,
comparing the benefits and drawbacks of selected approaches.
The next phase of this project focuses on the definition of a vehicle hetero-
geneous communication model to solve the network selection problem. The
main challenges and opportunities toward the interoperability of different
RATs were defined in Chapter 3. The primary objective during this phase
is to define the vehicular communication technique that can select the most
suitable RAT and trigger the VHO in such a way to minimize latency and the
handover impact, especially the packet loss and the ping-pong effect. The
proposed scheme should make an adequate choice of target technology and
provide a seamless handover. It can improve the available state of the art
methods or define new one.
In the next phase, the proposed approach should be implemented and
evaluated in a simulation environment. The objective of the simulation is
the initial analysis of the proposed solution. These simulations require re-
alistic traffic and network behaviour modelling in order to collect necessary
data from the simulation. Therefore, an overview of existing simulation tools
44
should be made before embarking on this project. The simulation tool should
model traffic scenarios and communication amongst the traffic participants.
It should be capable of simulating vehicular HetNet and support different
RATs. Consequently, a realistic simulation scenario should be defined and
evaluated as part of the following study as well. To evaluate the performance
of the proposed approach, several metrics should be defined and recorded dur-
ing the simulation runs. All these simulation outputs need to be thoroughly
investigated before final design decisions. At this point, another paper should
be published where detailed simulation results are presented and discussed.
After the simulation, experimental studies should be conducted to demon-
strate the applicability of the proposed approach in real-world environments.
First, the study of vehicular communication devices followed by hardware
selection and configuration should be done. This phase will be dedicated
to research on how to integrate our solution with existing equipment. The
test setup and test procedure should be first validated and calibrated in the
laboratory before moving to perform the field tests. The real-world deploy-
ment is expected to be done in the context of the UMinho-Bosch research
project, using the provided facilities (i.e. vehicles, OBUs, test field and other
required equipment). At this point, another paper should be published with
the results obtained. The work plan timeline is detailed in Figure 5.1. To
achieve the primary goal the tasks for work plan can be defined as listed
below:
1. Study the state of the art:
i Study the V2X development state of the art.
ii Investigation of the usefulness and applicability of multi-RAT us-
age for V2X.
iii Research and acquisition of knowledge from articles and previous
work on the theme.
iv Research studies on various HetNet aspects including architecture,
communication standards and applications;
2. Design:
i Requirements survey for multi-RAT vehicle network.
ii Selection of the potential approach form existing solutions.
iii Identification of the possible difficulties and the ways for solving
iv Design algorithms/techniques for V2X communication in a Het-
Net environment.
45
3. Simulation:
i Survey the existent V2X simulation tools/frameworks that allow
to simulate vehicle mobility and heterogeneous communication.
ii Implementation of the additional models and proposed algorithms
within the selected simulation environment.
iii Define and implement simulation scenarios.
iv Test of the proposed approach by means of simulation.
4. Real-world testing:
i Study and experiences with vehicular communication devices.
ii Hardware selection and configuration.
iii Design and build a prototype of the proposed solution.
iv Laboratory environment validation.
v Real world evaluation and tests.
5. Evaluation and documentation:
i Obtain and discuss the results from real tests and simulations.
ii Comparison with other state of the art solutions
iii Write and publish scientific documents that report the achieved
work.
iv Write a complete thesis document.
The dissertation writing task is expected to be conducted in parallel with
the others during the entire period of development of the thesis creates a
document that covers the content addressed at all stages and has, as a final
result, the document of the thesis.
46
Figure 5.1: Work Plan Timeline (divided by Quarters)
47
Chapter 6
Conclusions
V2X allows vehicles to communicate with other road users and infrastruc-
ture for enhancing road safety and mobility. The unique characteristics of
a vehicle communication environment are high mobility, low link availability
and variable network density. Recently, with advances in user equipment
and core networks the use of multi-technology enabled vehicular communica-
tion has attracted significant attention. A HetNet approach can combine the
benefits of each RAT to result in a more robust solution and flexibly react
to the highly dynamic vehicular environment.
The new concept called IoV arises to fulfill the challenging performance
requirements posed by emerging ITS applications and use cases. IoV com-
bines several networks to allow communication of different devices in a het-
erogeneous environment. These connections should be robust and efficient to
gather information and deliver it to all road users using more than one radio
interface. The process of selecting the most suitable network technology is
essential to maintain satisfactory connectivity in heterogeneous V2X environ-
ments. Since IoV is evolving as a global vehicular HetNet, a design of efficient
seamless handover procedure is important to maintain continuous connectiv-
ity with the high traffic load, low latency and high reliability requirements of
ITS services. For this thesis-planning document, we first provide an overview
of V2X heterogeneous communication including architecture, different RATs,
applications, handover process and equipment. Next, we surveyed novel ap-
proaches, highlighted key technical challenges and specified the opportunities
toward integration of wireless networks for vehicular connectivity. We have
studied the opportunities and challenges of applying the emerging solutions
to a vehicular HetNet based on multi-RAT. A taxonomy of recent advances
and open research directions is also provided. After this overview, we can
conclude that it is challenging to select appropriate RAT for a specific client
application while efficiently exploiting all the radio resources of HetNet.
48
Using the HetNet will lead to a significant enhancement and more effi-
ciency in vehicular communications. Thus, the interoperability of different
RATs remains an important area of research. We believe that in the next
few years addressing such topics will be a promising field of study as it could
overcome the shortcomings of traditional vehicular networks. The primary
contribution of this work is to provide insight into main design aspects of
hybrid V2X and define the research direction, considering the latest techno-
logical developments.
49
Bibliography
[1] Emmanuel Ndashimye, Sayan K. Ray, Nurul I. Sarkar, and Jairo A.
Guti´errez. Vehicle-to-infrastructure communication over multi-tier het-
erogeneous networks: A survey. Computer Networks, 112:144–166, 2017.
[2] Shanzhi Chen, Jinling Hu, Yan Shi, Li Zhao, and Wen Li. A Vision
of C-V2X: Technologies, Field Testing, and Challenges with Chinese
Development. IEEE Internet of Things Journal, 7(5):3872–3881, 2020.
[3] Carlos Renato Storck and Fatima Duarte-Figueiredo. A Survey of 5G
Technology Evolution, Standards, and Infrastructure Associated with
Vehicle-to-Everything Communications by Internet of Vehicles. IEEE
Access, pages 1–1, 2020.
[4] Karolis Kiela, Vaidotas Barzdenas, Marijan Jurgo, Vytautas Macaitis,
Justas Rafanavicius, Aleksandr Vasjanov, Leonid Kladovscikov, and Ro-
mualdas Navickas. Review of V2X-IoT standards and frameworks for
ITS applications, 2020.
[5] Ji Lianghai, Andreas Weinand, Bin Han, and Hans D. Schotten. Multi-
RATs support to improve V2X communication. In IEEE Wireless Com-
munications and Networking Conference, WCNC, volume 2018-April,
pages 1–6, 2018.
[6] Omprakash Kaiwartya, Abdul Hanan Abdullah, Yue Cao, Ayman Al-
tameem, Mukesh Prasad, Chin Teng Lin, and Xiulei Liu. Internet of Ve-
hicles: Motivation, Layered Architecture, Network Model, Challenges,
and Future Aspects. IEEE Access, 4(c):5356–5373, 2016.
[7] Haibo Zhou, Wenchao Xu, Jiacheng Chen, and Wei Wang. Evolution-
ary V2X Technologies Toward the Internet of Vehicles: Challenges and
Opportunities. Proceedings of the IEEE, 108(2):308–323, 2020.
50
[8] Pranav Kumar Singh, Sunit Kumar Nandi, and Sukumar Nandi. A
tutorial survey on vehicular communication state of the art, and future
research directions. Vehicular Communications, 18:100164, 2019.
[9] Gaurang Naik, Biplav Choudhury, and Jung Min Park. IEEE 802.11bd
5G NR V2X: Evolution of Radio Access Technologies for V2X Commu-
nications. IEEE Access, 7:70169–70184, 2019.
[10] Abdennour Zekri and Weijia Jia. Heterogeneous vehicular communica-
tions: A comprehensive study, 2018.
[11] Alessandro Bazzi, Giammarco Cecchini, Michele Menarini, Barbara M.
Masini, and Alberto Zanella. Survey and perspectives of vehicular Wi-Fi
versus sidelink cellular-V2X in the 5G era. Future Internet, 11(6), 2019.
[12] Shao Yu Lien, Der Jiunn Deng, Chun Cheng Lin, Hua Lung Tsai, Tao
Chen, Chao Guo, and Shin Ming Cheng. 3GPP NR Sidelink Transmis-
sions Toward 5G V2X. IEEE Access, 8:35368–35382, 2020.
[13] Sadip Midya, Koushik Majumder, Asmita Roy, and Debashis De. Ver-
tical handoff mechanisms in VANET : A survey. ACM International
Conference Proceeding Series, 04-05-Marc, 2016.
[14] Shimaa A. Abdel Hakeem, Anar A. Hady, and Hyung Won Kim. 5G-
V2X: standardization, architecture, use cases, network-slicing, and edge-
computing. Wireless Networks, 26(8):6015–6041, 2020.
[15] https://www.qualcomm.com/, 2020. [Online; accessed November-2020].
[16] Waqar Anwar, Andreas Trasl, Norman Franchi, and Gerhard Fettweis.
On the Reliability of NR-V2X and IEEE 802.11bd. In IEEE Interna-
tional Symposium on Personal, Indoor and Mobile Radio Communica-
tions, PIMRC, volume 2019-Septe, 2019.
[17] https://www.auto-talks.com/, 2020. [Online; accessed October-2020].
[18] Zeeshan Hameed Mir, Jamal Toutouh, Fethi Filali, and Young-Bae Ko.
Enabling DSRC and C-V2X Integrated Hybrid Vehicular Networks: Ar-
chitecture and Protocol. IEEE Access, 8:180909–180927, 2020.
[19] Kan Zheng, Qiang Zheng, Periklis Chatzimisios, Wei Xiang, and Yiqing
Zhou. Heterogeneous Vehicular Networking: A Survey on Architecture,
Challenges, and Solutions. IEEE Communications Surveys and Tutori-
als, 17(4):2377–2396, oct 2015.
51
[20] U Kumaran. Vertical Handover in Vehicular Ad-hoc Networks – A Sur-
vey. International Journal of Latest Trends in Engineering and Tech-
nology (IJLTET) - not index, 3(4):132–138, 2014.
[21] Jos´e Manuel Lozano Dom´ınguez and Tom´as Jes´us Mateo Sanguino. Re-
view on V2X, I2X, and P2X communications and their applications: A
comprehensive analysis over time. Sensors (Switzerland), 19(12), jun
2019.
[22] Fabio Arena and Giovanni Pau. An overview of vehicular communica-
tions. Future Internet, 11(2), 2019.
[23] Pranav Singh. Seamless V2I Communication in HetNet: State-of-the-
Art and Future Research Directions. Connected Vehicles in the Internet
of Things, 2020.
[24] Johann M´arquez-Barja, Carlos T. Calafate, Juan Carlos Cano, and
Pietro Manzoni. An overview of vertical handover techniques: Algo-
rithms, protocols and tools, 2011.
[25] Mir. QoS-Aware Radio Access Technology (RAT) Selection in Hybrid
Vehicular Networks. Communication Technologies for Vehicles, 2015.
[26] Shumin Bi, Cailian Chen, Rong Du, and Xinping Guan. Proper han-
dover between VANET and cellular network improves internet access.
IEEE Vehicular Technology Conference, 2014.
[27] R. Jacob, N. Franchi, and G. Fettweis. Hybrid V2X Communications:
Multi-RAT as Enabler for Connected Autonomous Driving. In IEEE
International Symposium on Personal, Indoor and Mobile Radio Com-
munications, PIMRC, volume 2018-Septe, pages 1370–1376, 2018.
[28] Richard Jacob, Waqar Anwar, Gerhard Fettweis, and Joshwa Pohlmann.
Exploiting Multi-RAT diversity in vehicular ad-hoc networks to improve
reliability of cooperative automated driving applications. IEEE Vehic-
ular Technology Conference, 2019-Septe:1–7, 2019.
[29] Miguel Sepulcre and Javier Gozalvez. Context-aware heterogeneous V2X
communications for connected vehicles. Computer Networks, 136:13–21,
2018.
[30] Miguel Sepulcre and Javier Gozalvez. Heterogeneous V2V Communica-
tions in Multi-Link and Multi-RAT Vehicular Networks. IEEE Trans-
actions on Mobile Computing, pages 1–1, 2019.
52
[31] Johann M. Marquez-Barja, Hamed Ahmadi, Sergio M. Tornell, Car-
los T. Calafate, Juan Carlos Cano, Pietro Manzoni, and Luiz A. DaSilva.
Breaking the vehicular wireless communications barriers: Vertical han-
dover techniques for heterogeneous networks. IEEE Transactions on
Vehicular Technology, 64(12):5878–5890, 2015.
[32] He Wei Yu and Biao Zhang. A Hybrid MADM Algorithm Based on At-
tribute Weight and Utility Value for Heterogeneous Network Selection.
Journal of Network and Systems Management, 27(3):756–783, 2019.
[33] Rodolfo I. Meneguette, Luiz F. Bittencourt, and Edmundo R.M.
Madeira. A Seamless FlowMobility Management Architecture for Ve-
hicular Communication Networks. Lecture Notes of the Institute for
Computer Sciences, Social-Informatics and Telecommunications Engi-
neering, LNICST, 58(2):42–56, 2013.
[34] Sandra C´espedes and Xuemin Sherman Shen. On Achieving Seamless
IP Communications in Heterogeneous Vehicular Networks. IEEE Trans-
actions on Intelligent Transportation Systems, 16(6):3223–3237, 2015.
[35] Jorge Dias, Andre Cardote, Filipe Neves, Susana Sargento, and Arnaldo
Oliveira. Seamless horizontal and vertical mobility in VANET. IEEE
Vehicular Networking Conference, VNC, pages 226–233, 2012.
[36] Pankaj Dahiya and Suman Deswal. Performance Analysis of Advanced
FMIPv6 For Handover Management In VANET. International Jour-
nal of Innovative Research in Science, Engineering and Technology,
3(6):13572–13578, 2014.
[37] Livinus Tuyisenge, Marwane Ayaida, Samir Tohme, and Lissan Ed-
dine Afilal. A mobile internal vertical handover mechanism for dis-
tributed mobility management in VANETs. Vehicular Communications,
26:100277, 2020.
[38] Muayad Khalil Murtadha, Nor Kamariah Noordin, Borhanuddin Mohd
Ali, and Fazirulhisyam Hashim. Fully distributed mobility management
scheme for future heterogeneous wireless networks. 2015 IEEE 12th
Malaysia International Conference on Communications, MICC 2015,
(Micc):270–275, 2016.
[39] Nigel Williams, Prashan Abeysekera, Nathan Dyer, Hai Vu, and
Grenville Armitage. Multipath TCP in Vehicular to Infrastructure Com-
munications. Grenville Armitage Centre for Advanced Internet Archi-
tectures Tech. Reports A. 140828, (August):15, 2014.
53
[40] Jorge Mena, Peter Bankole, and Mario Gerla. Multipath TCP on a
VANET: A Performance Study. SIGMETRICS 2017 Abstracts - Pro-
ceedings of the 2017 ACM SIGMETRICS / International Conference on
Measurement and Modeling of Computer Systems, pages 39–40, 2017.
[41] Pranav Kumar Singh, Sahil Sharma, Sunit Kumar Nandi, and Suku-
mar Nandi. Multipath TCP for V2I communication in SDN con-
trolled small cell deployment of smart city. Vehicular Communications,
15(November):1–15, 2019.
[42] Shiva Raj Pokhrel and Jinho Choi. Low-Delay Scheduling for Internet
of Vehicles: Load-Balanced Multipath Communication with FEC. IEEE
Transactions on Communications, 67(12):8489–8501, 2019.
[43] Liyang Sun, Guibin Tian, Guanyu Zhu, Yong Liu, Hang Shi, and David
Dai. Multipath IP Routing on End Devices: Motivation, Design, and
Performance. 2018 IFIP Networking Conference IFIP Networking and
Workshops, IFIP Networking 2018 - Proceedings, pages 181–189, 2018.
[44] Daiki Ishida, Daiki Nobayashi, and Takeshi Ikenaga. Path-selection
method based on the available bandwidth of interfaces. Proceedings -
2018 International Conference on Computational Science and Compu-
tational Intelligence, CSCI 2018, pages 815–818, 2018.
[45] Pedro Rezende, Somayeh Kianpisheh, Roch Glitho, and Edmundo
Madeira. An SDN-Based Framework for Routing Multi-Streams Trans-
port Traffic over Multipath Networks. IEEE International Conference
on Communications, 2019-May:1–6, 2019.
[46] Quentin De Coninck and Olivier Bonaventure. Multipath QUIC. pages
160–166, 2017.
[47] Jitendra Bhatia, Yash Modi, Sudeep Tanwar, and Madhuri Bhavsar.
Software defined vehicular networks: A comprehensive review. Interna-
tional Journal of Communication Systems, 32(12), aug 2019.
[48] Rodolfo Ipolito Meneguette. Software Defined Networks: Challenges
for SDN as an Infrastructure for Intelligent Transport Systems based on
Vehicular Networks. Proceedings - 16th Annual International Conference
on Distributed Computing in Sensor Systems, DCOSS 2020, pages 205–
212, 2020.
54
[49] Kan Zheng, Lu Hou, Hanlin Meng, Qiang Zheng, Ning Lu, and Lei
Lei. Soft-defined heterogeneous vehicular network: Architecture and
challenges. IEEE Network, 30(4):72–80, 2016.
[50] Zongjian He, Jiannong Cao, and Xuefeng Liu. SDVN: Enabling rapid
network innovation for heterogeneous vehicular communication. IEEE
Network, 30(4):10–15, 2016.
[51] Akshatha Nayak Manjeshwar, Arghyadip Roy, Pranav Jha, and Abhay
Karandikar. Control and management of multiple RATs in wireless
networks: An SDN approach. IEEE 5G World Forum, 5GWF 2019 -
Conference Proceedings, pages 596–601, 2019.
[52] Adnan Mahmood, Wei Emma Zhang, and Quan Z. Sheng. Software-
defined heterogeneous vehicular networking: The architectural design
and open challenges. Future Internet, 11(3), 2019.
[53] Ran Duo, Celimuge Wu, Tsutomu Yoshinaga, Jiefang Zhang, and
Yusheng Ji. SDN-based handover scheme in cellular/IEEE 802.11p hy-
brid vehicular networks. Sensors (Switzerland), 20(4):1–17, 2020.
[54] Kamal Deep Singh, Priyanka Rawat, and Jean Marie Bonnin. Cognitive
radio for vehicular ad hoc networks (CR-VANETs): Approaches and
challenges. Eurasip Journal on Wireless Communications and Network-
ing, 2014:1–22, 2014.
[55] Joy Eze, Sijing Zhang, Enjie Liu, and Elias Eze. Cognitive radio tech-
nology assisted vehicular ad-hoc networks (VANETs): Current status,
challenges, and research trends. ICAC 2017 - 2017 23rd IEEE Inter-
national Conference on Automation and Computing: Addressing Global
Challenges through Automation and Computing, (September):7–8, 2017.
[56] Nathalie Haziza, Mohamed Kassab, Raymond Knopp, J´erˆome H¨arri,
Florian Kaltenberger, Philippe Agostini, Marion Berbineau, Christophe
Gransart, Jo¨elle Besnier, Jacques Ehrlich, and Hasnaa Aniss. Multi-
technology vehicular cooperative system based on Software Defined Ra-
dio (SDR). Lecture Notes in Computer Science (including subseries
Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinfor-
matics), 7865 LNCS:84–95, 2013.
55
