ACCEPTED FROM OPEN CALL

6G Cellular Networks and Connected Autonomous Vehicles

Jianhua He, Kun Yang, and Hsiao-Hwa Chen

ABSTRACT intelligent surface (RIS), and artificial intelligence

(AI) [2, 4–6]. These revolution-natured technol-
With 5G mobile communication systems hav- ogies will drive the evolution of existing tech-
ing been commercially rolled out, research dis- nologies, such as advanced channel coding and
cussions on next generation mobile systems, that modulation, very large-scale antenna, spectrum
is, 6G, have started. On the other hand, vehicular sharing and full duplex, and so on.
technologies are also evolving rapidly, from con-
nected vehicles called V2X (vehicle to everything) C Connected autonomous vehicles (CAVs) [7]
is one of the critical vertical industries in 6G with
to autonomous vehicles to the combination of the its various demanding service qualities. There
two, that is, the networks of connected autono- are two levels of definitions for CAV. Basically,
mous vehicles (CAV). How fast the evolution of CAV can mean autonomous vehicles (AV) that
these two areas will go hand-in-hand is of great are connected to other vehicles and/or infra-
importance, which is the focus of this article. After structure. AVs are capable of sensing the driving
a brief overview of the technological evolution of environment and moving safely with little or no
V2X to CAV and 6G key technologies, this article human control. At an advanced level, CAV also
explores two complementary research directions, refers to the technologies and applications cen-
namely, 6G for CAVs versus CAVs for 6G. The tered around connected autonomous vehicles
former investigates how various 6G key enablers, that can collaborate with each other and infra-
such as THz, cell free communication and artifi- structure to achieve improved road safety and
cial intelligence (AI), can be utilized to provide efficiency compared to individual AVs without
CAV mission-critical services. The latter discusses cooperation. Initially, connected vehicles and
how CAVs can facilitate effective deployment and autonomous vehicles were developed in parallel,
operation of 6G systems. This article attempts to which are widely regarded as two of the most
investigate the interactions between the two tech- promising technologies for future transportation
nologies to spark more research efforts in these systems. However, they both have inherent short-
areas. comings. The combination of connected vehi-
cles and autonomous vehicles, thus giving rise to
INTRODUCTION so-called CAV, has attracted significant momen-
The mobile communication industry is one of very tum to tackle the transportation challenges. There
few industry sectors that have been fast-growing are many promising CAV applications, such as
for more than three decades. The upcoming 5G cooperated platooning, smart intersections, and
mobile networks promise to further change our cooperative perception, which can significantly
modern society and vertical industries with three improve road safety and efficiency, fuel consump-
identified services/use cases, including eMBB tion and congestion. To unleash the full potential
(enhanced mobile broadband), URLLC (ultra-re- of CAV, 6G needs to fulfill the following more
liable and low-latency communications), and stringent KPIs for the connectivity of vehicles,
mMTC (massive machine-type communications) including extremely high reliability: 99.999 per-
[1]. With 5G mobile systems being commer- cent; extremely low latency (0.1ms radio latency);
cially rolled out gradually, research discussions extremely massive instant access anytime and any-
on next generation mobile systems, that is, 6G, where; and extremely high throughput to support
have started [2–5]. 6G mobile network technol- a high volume of data transactions for full auto-
ogies are driven by the challenging demands of
emerging mobile applications, such as extended
>
mation. These KPIs pose significant challenges,
calling for new thinking and new communication
reality, industry 5.0 and digital twins, which go technologies that go beyond the current ones,
beyond 5G capacity. Though 6G is still in its con- such as Long Term Evolution (LTE) enabled C-V2X
ceptualization stage, some leading vendors have and 5G New Radio (NR) V2X [8, 9, 10]. We will
released initial drafts of technology-driven key discuss these issues in terms of two visional per-
performance indicators (KPIs) for 6G, including spectives as follows.
1 Tb/s peak data rate, 0.1 ms radio latency, max Vision 1: 6G will be a strong catalyst and
1 out of million outage reliability, 10 times more enabler for CAV’s mission critical services. 5G
energy efficiency, 20 years battery life time, and is targeted mainly for communications (either
100 devices per square meter density. There is a human-to-human or machine-to-machine). 6G
consensus on the potential enabling technologies will go beyond communications. Research has
for 6G, such as THz communications, integrated been conducted to utilize radio frequency (RF)
spatial-terrestrial networks (ISTN), reconfigurable to carry out object sensing and positioning. With
Digital Object Identifier:
Jianhua He and Kun Yang are with Essex University; Hsiao-Hwa Chen (corresponding author) is with National Cheng Kung University. 10.1109/MNET.011.2000541

IEEE Network • July/August 2021 0890-8044/20/$25.00 © 2020 IEEE 255

CHEN1_LAYOUT.indd 255
Authorized licensed use limited to: Thangal Kunju Musaliar College of Engineering. Downloaded on March 24,2023 at 14:09:25 UTC from IEEE Xplore. Restrictions 8/4/21
apply. 2:08 PM
 facilitates the implementation and deployment
of 6G in real life. Of course, the existing 4G and
emerging 5G technologies will also carry on to
support CAVs.

EVOLUTION OF V2X TO CAV

EVOLUTION OF V2X
Wireless connectivity is critical to CAV appli-
cations. Originated from the early research on
vehicular ad-hoc networks (VANETs) in the late
1990s, vehicle to everything (V2X) has been stud-
FIGURE 1. Architecture of CAV in 6G era. ied and standardized as the cornerstone of con-
nected vehicles. V2X is an umbrella of various
6G going toward a higher spectrum (such as communications technologies, covering vehicle
THz) and thus a shorter wavelength, a more pre- to vehicle (V2V), vehicle to infrastructure (V2I),
cise sensing and positioning resolution is theo- vehicle to pedestrian (V2P), and vehicle to net-
retically achievable. Higher frequency bands will work (V2N). The enabling V2X technology start-
enhance beamforming directionality and result ed with dedicated short range communications
in an increased data throughput. There is also (DSRC) with its spectrum in the 5.9 GHz band
an increasing trend of embedding intelligence allocated by the U.S. Federal Communication
(obtained via computation) into communication Committee (FCC) and Europe ETSI for intelligent
networks to cope with an increasing complexity transportation systems (ITS). DSRC can be used
of networks and network management. It is our to support many ITS applications such as toll col-
belief that 6G mobile systems facilitate missions lection, vehicle safety, in-vehicle entertainment,
of cross functions, combining communications and so on. The technology underpinning DSRC
(C), computation (C), positioning (P) and sensing is IEEE 802.11p, which is an amendment to the
(S), or CCPS for short, in order to satisfy service/ IEEE 802.11 standard to add wireless access in
application requirements while being deployed at vehicular environments with its particular focus
a large scale with cost effectiveness. A joint design on the PHY and MAC layers. Over the years, var-
of a 6G system multiple cross functions is likely to ious V2X configurations, both in academia and in
find a more efficient solution that is beneficial to industry, have been developed, including ITS-G5
all CCPS functions, which are also essential to a in Europe, and they have been based predomi-
CAV system to fulfill auto-driving. nantly on IEEE 802.11p.
Vision 2: CAVs will facilitate service provision- However, IEEE 802.11p suffers several limita-

5--
ing and operation of 6G. Roads are an integral tions due to its primary use of random channel
part of a modern city, in the same way as vehi- access, which include:
cles to a family or a society. Stationary roads and • Lack of QoS guarantee
mobile vehicles form an important part of our • Unbounded delay at physical channels
base infrastructure. A CAV system that is com- • Short-lived connectivity between vehicles
posed of smart roads, road side units (RSUs), and infrastructure
and vehicles can provide considerable resourc- • A need to deploy IEEE 802.11p infrastructure
es, physical space and services for communica- at a large scale, and so on.
tion, computing and intelligence. The unique CAV To overcome these limitations, the 3rd Genera-
features such as controlled mobility and ease of tion Partnership Project (3GPP) proposed cellular
deployment can effectively support the CCPS based V2X, namely C-V2X, which utilizes existing
functions of 6G systems through infrastructure cellular communication infrastructure for V2X [9,
extension, monitoring and maintaining 6G net- 10]. Unlike IEEE 802.11p, the cellular technolo-
works, to achieve 6G goals of ubiquitous wire- gies have inherent QoS mechanisms and are well
less intelligence and minimize network operation known for the merit of mobility management.
costs. With the expected tremendous investments C-V2X started with LTE, which is already commer-
in 6G and CAV infrastructures, the 6G com- cially deployed widely and used in our everyday
munications and CAV systems could be jointly life known as 4G.
designed, planned and operated with a much bet- Support of V2X services by 3GPP was spec-
ter reuse of system resources and services. ified in the Release 14 standards, which can be
This article aims to provide insightful discus- provided over the PC5 interface by sidelink trans-
sions on these two new research fields by bring- missions. The V2X transmissions are controlled
ing them together. This is one step further than by centralized scheduling via eNodeBs (LTE base
the latest ongoing research on 5G NR V2X and station) or gNBs (5G new radio base station) or
IEEE 802.11bd. Moreover, this article endeavors by distributed scheduling by vehicles. Centralized
to identify potential research directions of apply- broadcast scheduling has a high broadcast reli-
ing 6G to CAVs and vice versa. The relationship ability, but signaling overheads are very high due
between 6G and CAVs is illustrated in Fig. 1. As to frequent vehicle position update and resource
discussed earlier, the CAV KPIs and CCPS func- allocation. Distributed scheduling with autono-
tions guide and drive the design of a CAV sys- mous resource selection has a higher scalability.
tem, which consists of autonomous vehicles, RSU, The sidelink transmission for V2X was enhanced
smart road and also a control center. CCPS func- in 3GPP Release 15 with some new functions
tions and thus the CAV architecture are enabled such as transmission diversity, carrier aggregation
by 6G technologies such as THz, cell free and and higher quadrature amplitude modulation.
AI, and so on. In return, CAV infrastructure also 3GPP Release 16 specifies 5G NR based V2X

256 IEEE Network • July/August 2021

CHEN1_LAYOUT.indd 256
Authorized licensed use limited to: Thangal Kunju Musaliar College of Engineering. Downloaded on March 24,2023 at 14:09:25 UTC from IEEE Xplore. Restrictions 8/4/21
apply. 2:08 PM
 support with a few more new functions, including
the support for unicast and multicast, which are
particularly useful for CAV applications, enhanced
channel sensing, resource selection and QoS
management schemes, and congestion control [9,
10]. 3GPP Release 17 is in its planning stage, and
is expected to be completed in 2022, with further
enhancement on the sidelink efficiency, URLLC
and positioning, and use of relay and frequency
range 2 (FR2) over 6 GHz.
While C-V2X is developing rapidly in recent
years, IEEE is also catching up. A new Study FIGURE 2. A roadmap of V2X technological evolution, where 3GPP Release 17 is
Group called the IEEE 802.11 Next Generation in its planning stage and expected to be completed in 2021–2022.
V2X was formed in March 2018, which led to the
formation of IEEE Task Group 802.11bd (TGbd)
in January 2019. The ultimate goal of 802.11bd is tions, different strategies with a mixture of sensors
similar to that of 5G NR-V2X in that they both aim have been chosen for vehicles. For example,
to reduce latency, improve throughput, and pro- radar and lidar in addition to cameras are used
vide higher reliability (e.g., 99.999 percent). How- in the latest Waymo autonomous driving platform
ever, their design methodologies are different, (https://waymo.com/tech/). Tesla autonomous
namely, 802.11bd requires backward compatibili- vehicles uses cameras, radar and ultrasonic, but
ty on the same physical channel, whereas NR-V2X not lidar. The fusion of multiple local sensors
allows communications with its predecessor (LTE- could improve sensing and safety, but they are still
V2X) using a different radio channel [10]. While limited by line of sight (LOS) detection and the
the benefit of 802.11bd methodology is obvious, detection performance degrades with object dis-
it also introduces some huge challenges on its tance. Even though autonomous driving compa-
design and performance. A timeline on the evolu- nies are investing heavily in autonomous vehicles
tion of the V2X technologies is illustrated in Fig. 2. with increasingly powerful and more sensors, full
driving automation could still be out of reach due
EVOLUTION TO CAV to many factors, such as limitations of machine
Another timeline of CAV development is the learning algorithms and sensors, challenging driv-
increased level of automation in vehicles. Start- ing conditions and road emergencies, lack of
ing from basic advanced driving systems (ADS), redundancy on sensor safety and infrastructure
such as forward collision warning and automatic support. It turns out that connected vehicle is an
electronic braking, impressive CAV milestones excellent complementary technology to autono-
have been achieved with robotic taxi services mous vehicles and is widely regarded as an inte-
offered in the United States and China. The robot- gral part of fully autonomous vehicles. CAV as
ic taxis can drive autonomously under certain a combination of both can address most of the
conditions, which is at Level 4 of the Automated aforementioned challenges faced by autonomous
Driving standard set up by the Society of Auto- vehicles alone on the road to full automation.
motive Engineers (SAE) [11]. Many traditional car Furthermore, apart from enhancing driving safety
manufacturers and IT companies, such as Audi, by cooperative perception and cooperative driv-
Mercedes Benz and Google, are now striving to ing, CAV can enable many new road efficiency
achieve full automation (Level 5 of the SAE auto- applications, such as cooperative platooning and
mated driving standard). remote driving [7]. CAV has gained huge momen-
According to Mobileye, the largest ADS com- tum from the automotive and telecommunication
pany, there are three technological pillars for industries, academia and public authorities in the
autonomous driving, namely sensing, driving, last several years. A brief summary of sensors and
and mapping. The task of sensing is to build an connectivity used in CAV and their features is pre-
accurate environment model with 360 degrees sented in Table.1.
awareness, for example, detection of obstacles
and road signs. Various sensors have been uti-
lized in autonomous vehicles, including cameras, CAinclude
6G SUPPORT FOR CAV
typical CAV system, as shown in Fig. 3, may
Emplain key component
ultrasound, lidar, and short-range and long-range key components of CAVs on roads, RSUs CAUs in 6G
radars. All these sensors have their own strengths (equipped with communication, computing and of
and shortcomings. For example, cameras have traffic control devices), smart roads with intelligent
high resolution, long range detection, and capabil-
ity of recognizing road signs and traffic lights, but
materials and sensors, and transport control cen-
ters. The RSUs will play a critical role in collabo- X common with
they have problems when working in poor light- rative mobility and computing. In the meanwhile,

diagram
ing and weather conditions. Radar has the bene- unmanned aerial vehicles (UAVs) are applied in
fits of detecting objects with distance and speed, many scenarios to supplement on-road vehicles.
being robust against poor lighting and weather Connected unmanned aerial vehicles (CUAVs)
conditions, but has the shortcomings of low reso- are also regarded as a part of a CAV system. The
lution and is not able to recognize the shapes and new functions such as CCPS that could be offered
heights of the detected objects. Lidar has a wide by 6G systems can bring in significant benefits
field of vision and accurate distance estimation, to CAV from both connectivity and computing
but has poor resolution and is affected by bad aspects, also assisted by key technologies which
weather conditions. exist in 5G networks (such as mmWave, massive
To enable safe and reliable driving under vari- MIMO, network virtualization function or NFV,
ous challenging road, weather and lighting condi- and software defined networks or SDNs).
7
IEEE Network • July/August 2021 257

CHEN1_LAYOUT.indd 257
Authorized licensed use limited to: Thangal Kunju Musaliar College of Engineering. Downloaded on March 24,2023 at 14:09:25 UTC from IEEE Xplore. Restrictions 8/4/21
apply. 2:08 PM
I
Sensors Strength Shortcomings Distance (m) Data rate (Mb/s) No. of sensors Computation

Resolution, rich features, low cost, long No depth, poor weather-proof, LOS
Camera 250 100 10 High
range detection

Ultrasound Reliable detection Short distance, low resolution 10 0.01 10 Very low

Radar Resilience, depth and speed Low resolution, no height detection 300 10 5 Low

Lidar 3D detection, long range, resilience LOS, affected by poor weather 250 10 3 Medium

V2X NLOS, long range, resilience Rely on input from other vehicles 1000 10 1 Very low
TABLE 1. Comparison of CAV sensors (LOS: line of sight).

technologies relevant to CAVs will be discussed in

the sequel.
uTHZ FOR CAVS
Exploitation of the spectrum in the THz regime
is vital to achieve 6G KPI of 1 Tb/s data rate [2,
4]. The THz band usually refers to the frequency
band in the range of 100 GHz to 1 THz with the
corresponding wavelength of 30 mm to 3 mm.
The rich frequency resource available at the THz
band can provide a large bandwidth (up to 10s
GHz) and enable high speed links needed for
data intensive communications between CAVs
and between CAVs and infrastructure. Many
advanced CAV applications can be supported,
FIGURE 3. CAVs in 6G communications.
- including cooperative perception with raw sensor

explain
data, mobile edge intelligence and remote driv
ing. In addition to the communication benefits,
KEY 6G ENABLING TECHNOLOGIES ~ 6G systems operating on the THz band hold great
6G is envisioned to provide unprecedented potential on positioning, sensing and 3D imaging
capacity, sub-millisecond latency and ubiquitous [2]. As THz signals have short wavelength, the

Bribly
coverage [2]. It not only targets Tb/s data rates, antenna size and separation distance of antenna
but also inherently supports a wide range of novel elements can be reduced radically, which allows
scenarios and applications that combine agility, a large antenna array to be installed in mobile
reliability, ultra-low latency, and energy efficiency. devices and base stations. For example, more
There are already worldwide 6G research initia- than 200 antenna elements can be fit into an area
tives and programs, in which various technologies of 1 square cm at 300 GHz. Narrow beams can
and roadmaps have been proposed. In July 2018 be generated with a large antenna array, which
ITU formed a research group on 2030 network enables precision positioning for mobile devic-
technologies. In 2018, South Korea wireless tele- es with an outdoor location error of less than 1
communications operator SK Telecom proposed a meter. The high carrier frequencies of THz can
6G technical roadmap of THz, cell free networks, also enable radio frequency based sensing, pro-
and airborne wireless platforms. It agreed with viding accurate position and object detection. The
Samsung to work on joint 6G evolution technolo- positioning and sensing are highly appealing to
gies in 2019. China started 6G research initiatives CAVs. Traditional GPS satellite based positioning
to define Beyond 5G vision and requirements in methods suffer low location precision and limit-
2018. Huawei also announced its research study ed performance in urban environments. On the
of 6G technologies in 2019. In 2018, the FCC other hand, high definition mapping is not highly
envisioned 6G as featuring THz networks with responsive to the changes of roads such as road
multiple simultaneous beams of data transfer, construction and maintenance. 6G networks with
requiring unprecedented network densification. THz can provide a low cost, precision and reliable
The University of Oulu started the Finnish 6G alternative to positioning and sensing for CAVs,
Flagship program and organized the first 6G wire- which is extremely important to realize the full
less summit in 2019. They published a 6G white automation level. However, there will be major
paper [2], which identified key drivers, research challenges in designing a THz antenna/transceiv-
requirements and challenges. W. Saad et al. [5] er, increasing communication range, and dealing
discussed the 6G vision and technologies in a with high CAV mobility.

GRANT FREE AND NON-ORTHOGONAL MULTIPLE ACCESS (NOMA) ~

much broader sense. According to the report-
ed 6G programs and technical roadmaps, THz,
reconfigurable intelligent surface (RIS), ISTN, AI 6G systems are expected to meet the unprece-
and distributed computing have been tipped to dented requirement of supporting a massive num-
be the most promising candidate technologies ber of IoT devices while attaining ultra-reliability
for 6G. B. Letaief et al. [6] painted a 6G roadmap and low latency. A key enabling and candidate
with particular emphasis on the empowerment technology for 6G to meet the requirement is
of AI on 6G. Figure 3 illustrates how various 6G NOMA, which was proposed for 5G but has not
technologies can be integrated into CAV architec- yet been adopted by 3GPP. It allows multiple
ture and support CAVs. Some of these candidate users to utilize non-orthogonal resources concur-

258 IEEE Network • July/August 2021

CHEN1_LAYOUT.indd 258
Authorized licensed use limited to: Thangal Kunju Musaliar College of Engineering. Downloaded on March 24,2023 at 14:09:25 UTC from IEEE Xplore. Restrictions 8/4/21
apply. 2:08 PM
 rently for both random access and multiplexing. Applications Latency (ms) Data rate (Mb/s) Reliability Range (m)
NOMA can be achieved in the power domain,
code domain and pattern domain. NOMA has Cooperative awareness 100 0.01 0.9 300
been successfully applied to grant free access
Platooning 10~30 0.01~50 90 %~99.99 % 100~500
approaches [12, 13], which support massive con-
nectivity and achieve performance close to sched- Advanced driving 3~100 0.05~50 99.99 %~99.999 % 100~500
uled access schemes. Compared to orthogonal
multiple access, NOMA also has a superior spec- Extended sensors 3~100 25~1000 90 %~99.999 % 50~1000
trum efficiency. Remote driving 5~20 1~25 99.999 % 50~1000
NOMA is complementary to other 6G candi-
date technologies such as THz communication Edge driving 3~10 10~1000 99 %~99.999 % 50~1000
and can be used for both V2V and V2I communi- TABLE 2. Representative CAV applications enabled by 5G and 6G cellular net-
cations. It can be adopted in an enhancement to works.
the 5G V2X technologies and bring in great ben-
efits to CAV applications. CAVs usually operate
with high vehicle density and mobility, long com- ARTIFICIAL INTELLIGENCE (AI) AND EDGE INTELLIGENCE (EI) ~
munication range and heavy traffic conditions. Driven by recent breakthroughs in deep learning
NOMA could contribute with its capability of sup- and its successful applications in many areas such
porting massive connectivity and superior spec- as computer vision and natural language process-
trum efficiency. For example, sparse code multiple ing, there has been growing interest in the appli-
access (SCMA) can be applied to data resource cation of AI in mobile networks. It is envisioned
reservation in random access channels and data that AI will play a vital role in 6G and be applied
packet transmissions over data channels in the to many applications, such as communications
3GPP V2V protocols [12]. Research works have and networking, resource management, network
been reported on the applications of NOMA to control and automation [2, 6]. In addition, 6G is
V2X, which are focused mainly on V2I links and expected to provide edge intelligence (EI) ser-
centralized resource allocation [14]. Technical vices through mobile edge computing (MEC) to
challenges exist, including the design of practical IoT devices. CAVs rely heavily on the use of AI,
and efficient NOMA based grant free access and for example, in the environment perception for
multiplexing schemes for CAV applications, espe- accident avoidance, high-definition map for nav-
cially in distributed V2V networks scenarios. igation, and autonomous driving decision mak-
ing. The heavy computing loads of autonomous
CELL FREE COMMUNICATIONS w driving related applications can be offloaded to
To satisfy the need to support higher data rates in the MEC stations via high speed 6G links. For
6G networks, exploitation of mmWave and THz example, CAVs can forward sensor data to MECs,
frequency bands will be necessary. The cells oper- which process the data and aggregate sensing
ating in these frequency bands are small with a outcomes from other CAVs, then make real-time
radius of a few tens meters, and the communica- driving decisions for the CAVs or simply return
tion is subject to large path loss and signal block- the sensing outcomes such as obstacles or haz-
ing by obstacles. mmWave, THz and microwave ards on the roads. Distributed AI and federated
communication technologies will be integrated learning can also be supported with 6G MEC for
with co-existing multiple scales of cells, leading CAV cooperative learning and model training to
to frequent handovers in such multiple scale net- preserve user privacy.
works. It is more challenging for CAVs due to the Apart from the above discussed technologies,
strict requirements of CAVs on communication many others, such as QoS control, network slic-
and safety. ing and block-chain, could also be vital to ensure
A promising approach to tackle the mobility the strict requirements of CAV mission-critical
problem in 6G multi-scale networks is cell free services. Table 2 illustrates representative CAV
communication technology [15]. In a cell free applications that can be supported by 5G and
communication system, a large number of access 6G cellular networks, which provide various QoS
points (each with one or a few antennas) are performance metrics [8]. Apart from the basic
deployed and they cooperate via a backhaul net- applications of cooperative context awareness,
work and a central processing station to serve all 6G networks can support advanced CAV appli-
mobile users distributed over a wide area. In such cations, such as remote driving, edge driving and
systems, there are no cells or cell boundaries. The digital twins of CAVs [8]. Remote driving refers
mobile users move seamlessly within the heteroge- to the CAV applications, where the vehicles are
neous networks and receive the optimized service controlled remotely by human or machine drivers.
qualities. CAV applications can benefit from the Edge driving means autonomous vehicles are con-
cooperative communication technologies to avoid trolled by the machine drivers from the nearby
the aforementioned handover problems. The cell CAVs or edge computing stations. It is noted that
free networks also offer potential high precision for a specific CAV application, the requirement
positioning and sensing services from THz tech- on a specific QoS performance metric may vary
nology or additional sensors. The integrated cell according to the automation level of the vehicles.
free communication, positioning and sensing will For example, for the advanced driving application,
boost the performance of CAV applications with a semi-automation vehicle may require a commu-
improved cooperative perception and positioning. nication data rate of 50 kb/s, while a full automa-
Major technical challenges include coordination tion vehicle may need a communication data rate
of transmissions from the access points with high of 50 Mb/s. With advanced 6G support, it is also
vehicle mobility and radio resource management possible to create digital twins of CAVs, which will
by the central processing unit. collect CAV sensor data, build predictive models

IEEE Network • July/August 2021 259

CHEN1_LAYOUT.indd 259
Authorized licensed use limited to: Thangal Kunju Musaliar College of Engineering. Downloaded on March 24,2023 at 14:09:25 UTC from IEEE Xplore. Restrictions 8/4/21
apply. 2:08 PM
 for the CAVs, provide edge computing services,
control and support the CAVs.

CAV ENHANCEMENTS IN 6G NETWORKS

While 6G will be a strong catalyst and enabler for
CAVs, CAV in return can provide a strong support
to 6G in the delivery of communication, network-
ing, computing and management services. It will
facilitate the 6G systems to achieve the visions
of ubiquitous wireless intelligence. In this article,
we envision a 6G network architecture as shown
in Fig. 4, which includes mobile core network,
space access network, edge clouds and servers,
edge access networks and eventually end devic-
es. The mobile core network plays a similar role
as the gNBs in 5G networks, to manage mobili-
ty, network and connections to the Internet. The
space access network provides connectivity to the FIGURE 4. CAVs in 6G network architecture.
users via satellites and unmanned aerial vehicles
(UAVs). Central and edge servers provide net-
working and computing services with the features MOBILE VEHICLE EDGE COMPUTING
of network slicing and NVF. The edge access A key feature of 6G is the provision of intelligent
network connects mobile users and wireless end services and applications. A cloud computing
devices to the mobile core network and edge continuum (3C) consisting of remote clouds and
servers. Apart from the support of mmWave and edge clouds will be used to deal with the edge
THz communications, such as beam alignment computing applications with diverse computing
and connection management with CAV sensing, performance requirements. Due to the mobili-
the CAV system can enhance the 6G networks ty of edge computing service users, computing
from the edge and space access aspects. and data migration will be challenging under the
3C paradigm. With increasing computing power,
EXTENSION OF 6G COMMUNICATION INFRASTRUCTURE which is needed for high level automation of
Dense 6G base stations will be deployed, which CAVs, CAVs can support 6G networks in provid-
demand tremendous investment and are not flexi- ing mobile computing services. The availability of
ble to respond to fast spatial and temporal chang- storage space provides an opportunity for CAVs
es of cellular network traffic patterns. In addition, to be equipped with more resources for larger
under emergency situations, such as natural disas- scale vehicle edge computing (VEC) services.
ters, the communication infrastructure may be They can be deployed according to the demands
damaged or not available. The surface CAVs and of 6G network operators or individuals, support-
connected UAVs (CUAVs) provide an excellent ing the other CAVs or other types of mobile users.
mobile platform to extend the 6G communication VECs with features, such as autonomous mobility
infrastructure, which can be converted to CAV and ease of deployment, represent a flexible and
mobile base stations (CBS) and deployed to the great enhancement to 6G MEC.
locations as needed to provide a wider coverage.
CBS provides a flexible and economic solution NETWORK PERFORMANCE MONITORING AND SENSING FOR 6G
to augment fixed 6G communication infrastruc- 6G networks will have sophisticated structure and
ture. The CBSs can be self-organizing and coop- require complex management and control. There is
erate with each other, deciding where and when a trend of increasing network automation and resil-
to provide communication services adaptively in ience from telecommunication equipment vendors
response to changing demands. While it is possible and network operators. CAV can contribute to the
that some CBSs are dedicated to providing com- automation of 6G networks. Performance mea-
munication services, most of them are expected to surement and monitoring are critical to achieve
have major duty of providing mobility services. intelligent control and automation, for which CAVs
While it is possible for the CAVs to provide and smart roads with intelligent sensors can con-
access service in moving, they are assumed to tribute with real time updates on service quality
provide the service when they are stationary and and surveillance of network infrastructure. The
play a supportive role by offloading a small part CAVs and smart roads can support the construc-
of traffic instead of a major part of traffic from tion and operation of a 6G RIS based smart com-
the legacy base stations. The supporting CAVs munication environment. They can be deployed on
will establish connections to the base stations for demand to measure network service quality, act as
wireless backhaul purpose. temporal base stations to enhance 6G communi-
Interestingly, road traffic and wireless network cations, detect and even repair problematic out-
traffic generate patterns, which are suitable for door 6G network components. With an increasing
the reuse of the CAVs in providing communica- penetration of CAVs they can provide an econom-
tion services. For example, at night time when ic alternative to slow and costly manual network
there is less traffic on roads but more wireless net- monitoring and enhance 6G automation.
work traffic, fewer CAVs are needed for transpor-
tation and more CAVs can be used as CBS. The CONCLUSIONS
utilization of the CAVs can be largely improved This article brought together two promising
and the 6G network operational costs can be research directions, that is, CAVs and 6G net-
reduced significantly. works, and discussed their potential interac-

260 IEEE Network • July/August 2021

CHEN1_LAYOUT.indd 260
Authorized licensed use limited to: Thangal Kunju Musaliar College of Engineering. Downloaded on March 24,2023 at 14:09:25 UTC from IEEE Xplore. Restrictions 8/4/21
apply. 2:08 PM
 tions and mutual supports. After drawing out
a roadmap of the technical evolution of V2X It is our belief that the intersection of CAV systems and 6G networks will bring in significant innovations
to CAV and that of 6G key technologies, this and momentum to the development in both areas. A joint design of both may be an effective way forward
article explored two complementary directions
of future research, namely, 6G for CAVs and and such a consideration shall be taken as much as possible from the early design stage of each.
CAVs for 6G. Discussions have been made to
show how various 6G key enablers, such as [11] A. Qayyum et al., ”Securing Connected and Autonomous
THz, cell free communications and edge intel- Vehicles: Challenges Posed by Adversarial Machine Learn-
ligence, can be utilized to enable CAV’s mis- ing and the Way Forward,” IEEE Commun. Surveys and Tuto-
rials, vol. 22, no. 2, 2020, pp. 998–1026.
sion-critical services. Proposals are also made [12] A. C. Cirik et al., ”Toward the Standardization of Grant-Free
to illustrate how CAVs can be employed for a Operation and the Associated NOMA Strategies in 3GPP,” IEEE
more effective and efficient deployment and Commun. Standards Mag., vol. 3, no. 4, Dec. 2019, pp. 60–66.
operation of 6G systems. It is our belief that [13] S. Guo and X. Zhou, ”Robust Resource Allocation with
Imperfect Channel Estimation in NOMA-Based Heteroge-
the intersection of CAV systems and 6G net- neous Vehicular Networks,” IEEE Trans. Commun., vol. 67,
works will bring in significant innovations and no. 3, Mar. 2019, pp. 2321–32.
momentum to the development in both areas. [14] Di B. et al., ”Non-Orthogonal Multiple Access for High-Re-
A joint design of both may be an effective way liable and Low-Latency V2X Communications in 5G Sys-
tems,” IEEE JSAC, July 2017.
forward and such a consideration shall be taken [15] H. Q. Ngo et al., ”Cell-Free Massive MIMO Versus Small
as much as possible from the early design stage Cells,” IEEE Trans. Wireless Commun., vol. 16, no. 3, Mar.
of each, aiming to achieve an optimized inte- 2017, pp. 1834–50.
gral system that benefits both sectors.
BIOGRAPHIES
ACKNOWLEDGMENT JIANHUA HE received his Ph.D. degree from Nanyang Techno-
This work was supported in part by the European logical University. He is a reader at the University of Essex, U.K.
Union’s Horizon 2020 research and innovation His research interests include mobile networking and comput-
ing, Internet of Things, edge computing, CAV and machine
programme under the Marie Skodowska-Curie learning. He has published over 150 scientific papers in these
grant agreement No 824019; the National Natu- areas. He is the coordinator of the EU H2020 COSAFE project
ral Science Foundation of China (U1764263 and on cooperative CAV. He is a Senior IEEE Member. He served
61671186); and the Taiwan Ministry of Science as editor or guest editor for Frontiers in Future Transportation,
Future Internet, IEEE Access, and Wireless Communications and
& Technology (109-2221-E-006-175-MY3 and Mobile Computing. He is a TPC co-chair of MobiArch2020 and
109-2221-E-006-182-MY3). ICAIT2009, and General Chair of ICAIT and ICONI.

REFERENCES KUN YANG received his Ph.D. degree from the Department of
[1] J. G. Andrews et al., ”What Will 5G Be?” IEEE JSAC, vol. 32, Electronic and Electrical Engineering at University College Lon-
no. 6, June 2014, pp. 1065–82. don (UCL), U.K. He is currently a chair professor in the School
[2] M. Latva-aho and K. Leppnen (eds.), ”Key Drivers and of Computer Science and Electronic Engineering, University of
Research Challenges for 6G Ubiquitous Wireless Intelli- Essex, U.K., leading the Network Convergence Laboratory (NCL).
gence,” 6G Flagship, Sept. 2019. His main research interests include wireless networks and com-
[3] P. Yang et al., ”6G Wireless Communications: Vision and munications, future Internet and edge computing. He manages
Potential Techniques,” IEEE Network, vol. 33, no. 4, July/ research projects funded by UK EPSRC, EU FP7/H2020 and
Aug. 2019, pp. 70–75. industry. He has published 300+ papers and filed 10 patents. He
[4] Z. Zhang et al., ”6G Wireless Networks: Vision, Require- serves on the editorial boards of a number of IEEE journals.
ments, Architecture, and Key Technologies,” IEEE Vehic-
ular Technology Mag., vol. 14, no. 3, Sept. 2019, pp. HSIAO-HWA CHEN [S’89, M’91, SM’00, F’10] is currently a Dis-
28–41. tinguished Professor in the Department of Engineering Science,
[5] W. Saad, M. Bennis, and M. Chen, ”A Vision of 6G Wire- National Cheng Kung University, Taiwan. He obtained his B.Sc.
less Systems: Applications, Trends, Technologies, and Open and M.Sc. degrees from Zhejiang University, China, and a Ph.D.
Research Problems,” IEEE Network, 2020. degree from the University of Oulu, Finland, in 1982, 1985, and
[6] K. B. Letaief et al., ”The Roadmap to 6G: AI Empowered 1991, respectively. He has authored or co-authored over 400
Wireless Networks,” IEEE Commun. Mag., vol. 57, no. 8, technical papers in major international journals and conferences,
Aug. 2019. six books, and more than 10 book chapters in areas of communi-
[7] H. Abou-zeid et al., ”Cellular V2X Transmission for Connect- cations. He has served as the general chair, TPC chair, and sym-
ed and Autonomous Vehicles Standardization, Applications, posium chair for many international conferences. He has served
and Enabling Technologies,” IEEE Consumer Electronics or is serving as an editor or guest editor for numerous technical
Mag., vol. 8, no. 6, Nov. 2019, pp. 91–98. journals. He is the founding Editor-in-Chief of Wiley’s Security and
[8] 3GPP TS22.186, ”Enhancement of 3GPP support for V2X Communication Networks Journal. He is the recipient of the best
scenarios, v16.1, Dec. 2018. paper award in IEEE WCNC 2008 and the recipient of the IEEE
[9] S. Lien et al., ”3GPP NR Sidelink Transmissions Toward 5G 2016 Jack Neubauer Memorial Award. He served as the Editor-
V2X,” IEEE Access, vol. 8, Feb. 2020, pp. 35368–72. in-Chief for IEEE Wireless Communications from 2012 to 2015.
[10] G. Naik, B. Choudhury, and J. Park, ”IEEE 802.11bd & 5G He was an elected Member-at-Large of the IEEE Communications
NR V2X: Evolution of Radio Access Technologies for V2X Society from 2015 to 2016. He is serving as TPC Chair for IEEE
Communications,” IEEE Access, vol. 7, June 2019. Globecom 2019. He is a Fellow of IEEE, and a Fellow of IET.

IEEE Network • July/August 2021 261

CHEN1_LAYOUT.indd 261
Authorized licensed use limited to: Thangal Kunju Musaliar College of Engineering. Downloaded on March 24,2023 at 14:09:25 UTC from IEEE Xplore. Restrictions 8/4/21
apply. 2:08 PM