A New Era of Mobility: Exploring Digital Twin Applications in Autonomous Vehicular Systems

Abstract

Digital Twins (DTs) are virtual representations of physical objects or processes that can collect information from the real environment to represent, validate, and replicate the physical twin's present and future behavior. The DTs are becoming increasingly prevalent in a variety of fields, including manufacturing, automobiles, medicine, smart cities, and other related areas. In this paper, we presented a systematic reviews on DTs in the autonomous vehicular industry. We addressed DTs and their essential characteristics, emphasized on accurate data collection, real-time analytics, and efficient simulation capabilities, while highlighting their role in enhancing performance and reliability. Next, we explored the technical challenges and central technologies of DTs. We illustrated the comparison analysis of different methodologies that have been used for autonomous vehicles in smart cities. Finally, we addressed the application challenges and limitations of DTs in the autonomous vehicular industry.

Full text

A New Era of Mobility: Exploring Digital Twin
Applications in Autonomous Vehicular Systems
S M Mostaq Hossain
Dept. of Computer Science
Tennessee Tech University
Tennessee, USA
shossain42@tntech.edu
Sohag Kumar Saha
Dept. of Electrical &
Computer Engineering
Tennessee Tech University
Tennessee, USA
ssaha42@tntech.edu
Shampa Banik
Dept. of Computer Science
Tennessee Tech University
Tennessee, USA
sbanik42@tntech.edu
Trapa Banik
Dept. of Electrical &
Computer Engineering
Tennessee Tech University
Tennessee, USA
tbanik42@tntech.edu
Abstract—Digital Twins (DTs) are virtual representations of
physical objects or processes that can collect information from
the real environment to represent, validate, and replicate the
physical twin’s present and future behavior. The DTs are be-
coming increasingly prevalent in a variety of fields, including
manufacturing, automobiles, medicine, smart cities, and other
related areas. In this paper, we presented a systematic reviews
on DTs in the autonomous vehicular industry. We addressed DTs
and their essential characteristics, emphasized on accurate data
collection, real-time analytics, and efficient simulation capabili-
ties, while highlighting their role in enhancing performance and
reliability. Next, we explored the technical challenges and central
technologies of DTs. We illustrated the comparison analysis of
different methodologies that have been used for autonomous
vehicles in smart cities. Finally, we addressed the application
challenges and limitations of DTs in the autonomous vehicular
industry.
Index Terms—digital twin; vehicular network; smart vehicles;
autonomous driving; literature review; cyber-physical systems.
I. INTRODUCTION
Over the past decade, autonomous driving (AD) has grown
fast, transforming the transportation system in terms of safety
and efficiency [1]. As AVs proliferate, safety and dependability
are paramount in AV system development. The latest research
implies that AVs could considerably improve vehicle safety
[2], but this won’t be achievable until a fleet of AVs has tested
billions of kilometers in all weather situations. Running a
fleet of AVs and development infrastructures that use physical
testing data would take decades and tens of billions of dollars
to meet AV [3] safety targets.
Digital Twin (DT) technology has various uses, from real-
time remote monitoring and control in industry to risk as-
sessment in transportation to smart scheduling in smart cities,
therefore it has received a lot of attention recently. According
to Z. Hu et al. [4], figure 1 shows the major DT development
milestones. In a high-fidelity virtual environment, simulation-
based digital twins can speed up AV verification and save
development expenses [5]. Software must change for digital
twins. Digital twins can simulate various environmental and
traffic situations, avoiding the need for extensive physical
testing. Virtual, controllable testing of autonomous vehicles
could save several orders of magnitude in development time
and cost. We may have to wait months for significant snow to
test AVs on the road. A digital twin can build a road, simulate
a major snowstorm, and generate a lot of high-quality testing
data [6]. Our initial deployment of a digital twin system that
combines physical and digital twin testing has been successful,
but its flexibility allows for further improvement [7].
Vehicle dynamic simulators, including cruise control system
simulators, have been widely used in the automobile industry
for testing [8]. Aerospace has long used simulation. Testing
autonomous vehicles in virtual and controllable environments
could save development costs and time by orders of magnitude.
In the development of AD software , simulators have been
used extensively to test and evaluate the decision-making
module and path-planning module under various scenarios by
providing perception data (such as the position and moving
states of the ego vehicle and other traffic participants) [9]. This
strategy is easy and scalable, however it does not represent the
reality well, which causes problems. Simulation tests cannot
compare to physical AV software pipeline tests. This pipeline
includes sensing, localisation, decision-making, path planning,
and vehicle control. Physical examinations. Physical elements,
such as weather and lighting, also prevent investigations.
First, these simulators use virtual town maps instead of real-
world road testing [10]. Not simulated are road conditions. A
digital twin map is needed to evaluate AD functions like exit
or entry highway ramps that depend on road geometry and
traffic legislation. Simulators’ car and pedestrian animations
are pre-programmed. Simulations cannot replicate real traffic’s
intricate interactions. Junctions and aggressive driving are un-
judgable. AV software testing suffers from low-fidelity sensor
data. Depth mapping and ray casts replicate lidar sensors but
don’t account for reflection and diffusion. Simulations differ.
Unlike AV software development, automobile hardware de-
velopment accelerates with ”digital twin” physical simulation
tools like MATLAB and Modelica.
Figure 2 shows how our two-tiered framework creates a
linked vehicle digital twin system [11]. Virtual is above actual.
This system’s communication module is vital. Cellular data
transport powers this experiment. The physical layer of the
digital twin framework can represent all physical entities and
their interactions, such as automobiles and their parts, drivers
and passengers, road infrastructure, weather, other road users,
arXiv:2305.16158v1 [cs.NI] 9 May 2023

Fig. 1. History of digital twin technology
etc., defined on a global coordinate system and developing
over time. Sensors and actuators. The sensors may detect and
aggregate vehicle speed, driver gaze, and traffic light status at
various resolutions. The communication module analyzes data
online. The item or process, sensors, actuators, and computing
resources are shown in Figure 2 [13]. The digital world
comprises databases, data processing infrastructures, machine
learning, and the digital twin. Wi-Fi and Bluetooth protocols
and interfaces link them. Architecture monitoring requires
visualization.
The key contributions of this work are as follows:
•We investigated the overview and recent researchers on
the DTs concept specifically for autonomous vehicular
systems.
•The state-of-the-art research methodologies are discussed
based on current literature’s of DTs.
•The comparison of different methods, their role in tech-
nical development and limitations are stated.
The remaining sections of this paper is organized in the fol-
lowing manner: Section II provides a technological overview,
Section III presents a brief literature review, Section IV
outlines the methodologies used, and Section V presents a
comparative analysis. The remaining sections such as: Section
VI and Section VII, consists of the Discussion and Conclusion,
respectively.
II. AN OV ERV IE W OF D IGITAL T WI N TE CH NO LO GY F OR
SMART VEHICLES
Digital twins are used in the automotive industry to create
digital copies of vehicles. Data on car use and performance en-
ables for more personalized service and maintenance. Digital
copies can be model copies or networked systems [14]. Engi-
neers investigate AI before sending a car to the assembly line.
Simulation models can predict breakdowns and wear. Instead
of road testing and maintenance, autonomous vehicle digital
twins could save unforeseen costs. Digital twin technology
may imitate and improve many aspects of a smart electric
vehicle’s system, which has far-reaching effects [15]. Digital
twin vehicle modeling requires understanding the digital twin
environment.
Automakers employ digital twins to make digital duplicates
of automobiles. Car performance data allows for more tar-
geted service and maintenance. Model copies or networked
digital copies [14]. Engineers study AI before assembling
an automobile. Simulations forecast breakdowns and wear.
Autonomous vehicle digital twins could eliminate road testing
and maintenance. Digital twin technology can replicate and
improve a smart electric vehicle’s system. Digital twin vehicle
modeling requires digital twin environment knowledge.
III. LITERATURE REVIEW
In this section, we’ve presented a brief overview of the DT
concept for the autonomous vehicular industry. The following
discussion has included the literature overview regarding our
goal.
In the paper, [16] B. Yu, et al. shared real-world experi-
ences of the digital twins, a practical method for developing
autonomous driving (AD) systems that creates a complete,
accurate, and reliable model of the physical environment to
2

Fig. 2. General framework of the digital twin system for connected vehicles [11]
reduce the need for physical testing. Their main contributions
are:
•they have identified the limitations of conventional ap-
proaches to AD simulation and show how digital twins
can be used to overcome them.
•To begin, they synthesized their practical development
experience into three overarching concepts for the AD
digital twin system’s design.
•at the end, they described the AD digital twin system’s
structure and its components, including how real-world
mapping data is collected, sensor data is mimicked, and
traffic actors are synthesized.
The followings are the paper’s [17] unique contributions in
comparison to previous recent research on the validation of
coordination strategies:
•At the end, they described the AD digital twin system’s
structure and its components, including how real-world
mapping data is collected, sensor data is mimicked, and
traffic actors are synthesized.
•Under the digital twin paradigm, a working prototype has
been created. Time lag and precision of localization are
just two of the worrying metrics tracked.
•The system includes HMI devices. In 3D, the Hololens
controls automobiles. First-person perspective driving
simulators allow drivers full control.
The study [18] examined digital twin technology’s origins
and deployment phases. This research highlights digital twin
technologies like predictive mobility, autonomous motion con-
trol, driver assistance systems, vehicle health management
systems, battery management systems, intelligent charging,
vehicle power electronic converters, and electric power drive
systems. Barriers to adoption and important supporting tech-
nologies are also identified, which will aid future eco-friendly
and sustainable transportation endeavors.
The contribution of the paper [19] are follows:
•In densely populated locations with considerable traffic,
automobiles are increasingly being used as services due
to commercial expansions. This study suggested using
DT to enable CaaS.
•The suggested design includes the city, a middleware
to connect all entities, DTs models that run over the
middleware, and applications like car-sharing.
•The case study showed how the proposed concept might
be implemented and highlighted several key areas for
development in future efforts.
The paper [20] presented a Petri-net-based DT to simulate
the electric car development process from start to finish.
Real-time data sharing between the physical system and its
digital shadow can help make better, faster decisions. The
two systems can directly calculate and implement actions
to contain these facts, calculate new time plans, and inform
the user of optimistic, most likely, and pessimistic scenarios
for task delays. Our research will improve search algorithms
for non-computable solutions and optimize physical-digital
subsystem communication and interaction.
This study proposes a V2C-communicating linked car dig-
ital twin structure [21]. The vehicle’s driver-vehicle interface
(DVI) displays the cloud server’s advisory speed, letting
3

the driver control the vehicle. ADAS uses the digital twin
paradigm. The suggested digital twin structure is tested in real-
world traffic on three passenger vehicles in a cooperative ramp
merging case study.
In paper [22], the author’s plan is to create an unsupervised
prognosis and control platform tailored to electric propul-
sion drive systems (EPDS) performance estimation as a final
product. A number of subsidiary tasks and goals must be
formulated in order to accomplish this primary aim.
•Construct the DT of the energy system by creating
physical models of its constituent parts (motors, gener-
ators, gearboxes, bearings, etc.) and their corresponding
reduced models (testbed).
•Create a working prototype of the Virtual Sensors idea,
which is built on the existing DT concept.
•Make a system based on artificial intelligence that lets
the virtual sensors be used to control EPDS.
•Demonstrate the understanding of these ideas and how
they apply to achieve the aforementioned end goal using
the autonomous vehicles as a case study.
IV. METHODOLOGIES OF THE SELECTED PIECES OF
LI TE RATU RE
There have been several methodologies have been discussed
of the selected pieces of literature. Among them are discussed
below.
A. Three Principles architecture
The authors have created a digital twin system using a game
engine, which includes graphics and physics engines for 3D
modeling, image rendering, and physical simulation, allowing
for the representation of structural, physical, and behavioral
information in a virtual world [16]. In order to implement the
digital twin properties, three additional building blocks are
added on top of the game engine: It can be seen that 1) the
traffic controller is the logical twin, 2) sensor models are the
physical twin, and 3) the 3D digital twin map is the structural
twin.
B. Multi-vehicle Experiment Platform
Tsinghua is developing a CAV-focused digital twin system.
Even without real cars, this technology will allow multi-
vehicle tests. This study [18] advises using real and virtual
autos to get the desired outcome. A sand table testbed helps
small cars run smoothly. A game engine creates cyberspace
through full-element modeling. Cyberspace can show the sand
table’s real-time state. This research proposes a cloud vehicle
to replace smaller autos.
C. Car-as-a-Service Concept with DT
Pana offers Car-as-a-Service (CaaS) employing sensors, ac-
tuators, and radio devices [23]. Smart cities should use linked
automobiles for passenger service, the authors say. GNSS,
high-precision distance estimation, radio connectivity, environ-
mental sensors for measuring temperature, humidity, pressure,
etc., motion sensors like an Inertial Measurement Unit (IMU)
for measuring traffic flow, vehicle heading and roll, and road
quality, a centralized cloud processing unit, and social network
analytics are the backbone of CaaS [19]. CaaS may entail
carpooling. Users rent and share cars. This service cuts traffic
and saves occasional drivers. Archer [24] cites studies showing
that five to fifteen privately owned cars are replaced for
every shared car added to the fleet, assuming car-sharing
programs reduce car ownership. Ferrero [25] investigates car-
sharing. Technical and modeling studies have examined car-
sharing businesses. This service is hard to sell. European,
US, Japanese, Chinese, and Australian car-sharing schemes
are widespread. Mattia [26] expects 12 million users by 2021.
User settings [27] and driver monitoring will increase comfort
and safety in this car-sharing concept. [28] found that Drive
Monitoring Assistance System (DMAS) must monitor drivers’
attention levels for safe driving. Distraction and fatigue cause
most traffic accidents.
D. Petri nets and variations
Modeling, simulating, and tracking development Arc-
extended timed Petri nets will be used [20]. Formally, an
Ordinary Petri Net (OPN) is a five-tuple with these elements:
PN requires a finite number of locations (P= p1, p2,..., pnp),
transitions (T= t1, t2,..., tnt), vertices (V), and an empty
set (PT=) as their intersection. In the PN token distribution,
m0 is the initial token distribution, I and O are input and
output functions (PNs initial marking). Locations in a Petri net
network represent resources, governing conditions, transitions,
and arcs [29].
T-timed PNs implement transition time delays. They behave
like Ordinary PNs but can predict event durations, making
them better for simulation. Time-delayed PNs, or T-timed
PNs, are defined as TPN= ”P, T, I, O, m0, D,” where D
is a function of positive real integers. Using arc extensions
to activate or deactivate PN sections when certain criteria
are satisfied is critical for control. Standard, inhibitor, and
activator arcs (illustrated as dashed vectors; [30], [31]) are
used across the literature. Arc extensions boost the baseline
model’s simulation capacity by allowing more complex ideas
with fewer node connections.
E. Vehicle-to-Cloud Based Advanced Driver Assistance Sys-
tems
Authors create a two-layer digital twin for linked cars [21].
Cyber tops physical. Communication links two system frame-
work levels. Study communication is cellular. The physical
layer of the digital twin system, specified on a world coordi-
nate throughout time, may contain cars, components, drivers,
passengers, roadway infrastructure, meteorology, other road
users, etc. Sensors and actuators are key layers. Vehicle speed,
driver look, and traffic signal status can be detected by the
sensors. Cyberspace processes data from the communication
module.
Through the communication module, cyberworld beings and
processes become corporeal. ADAS-equipped people can drive
connected cars. Cyberworld actuation guidance guides the
4

TABLE I
COMPARISON ANALYSIS OF DIFFERENT DIGITAL TWIN SYSTEMS FOR AUTONOMOUS VEHICLE SYSTEM
Sl. Methodology Used Role of DT Technology Evaluation Future Aspects Year [Ref.]
1
DT-GM: SUMO’s innovative
microscopic simulation-based
Geneva highway digital twin.
DT-GM’s development process
is explained, highlighting
SUMO’s calibration features
that allow ongoing calibration
of running simulation scenarios.
Experimental findings show
that DT-GM accurately
represents traffic. ODPMS’s
current development enables
DT-GM’s low-latency
motorway response.
Usefulness of powerful
self-regulating adaptive
controllers to optimize
variable speed restrictions
in safety-critical
decisions using DT.
2023 [38]
2
Examined the IoDT’s
architecture, communication
modes, major aspects,
supporting technologies,
and recent prototypes.
We explore IoDT security
and privacy challenges
from seven perspectives—data,
authentication, communication,
privacy, trust, monetization,
and cyber-physical—and the
main obstacles to tackling them.
Comprehensive understanding
of IoDT working principles,
including its general architecture,
key characteristics, security or
privacy threats, and existing or
potential countermeasures.
Cloud-Edge-End Orchestrated,
Space-Air-Ground Integrated,
Interoperable and Regulatory,
Explainable AI-Empowered,
Information Bottleneck Based,
and Privacy-Aware of IoDT
2023 [39]
3
Developed a digital twin a
system using a game engine
for 3D modeling, picture
rendering, and physical
simulation. Structural,
physical, and logical twins
are introduced.
Three components implement
digital twin properties: The
3D digital twin map is
structural, sensor models are
physical, and the traffic
controller is logical.
The digital twin paradigm creates
an entire, comprehensive, exact,
and trustworthy representation of
the physical environment,
allowing for quick, low-cost
development iterations.
Digital twin cost and
efficiency can be
improved by optimizing
the computing system.
2022 [16]
4
The archetypal infrastructure
uses sand table, clone, and
cloud. A driving simulator,
Hololens, and a screen display
improve system engagement.
A platoon based test proves
the system can test many
cars simultaneously.
This research advises
employing digital twins
instead of real automobiles
for multi-vehicle studies.
The system is also modeled.
Virtual speed fluctuations are
smoother, and speed profiles
between the same type of
vehicles are comparable, but
those between different types
vary greatly. To follow a physical
vehicle, V2’s spacing variations
are unsteady, like small ones.
1) Impact of different
access modes on
experiment results
will be analyzed.
2) More dynamics
models of cloud
vehicles will be provided
2022 [17]
5
This article presents an
eclectic overview of the
smart car system by
discussing each
component in depth.
The study includes predictive
mobility, autonomous motion
control, driver aid systems,
vehicle health management
systems, battery management
systems, and electric power
drive systems.
1) Generating historical collision
time and speed data from case
scenarios. 2) Use real-time vehicle
data and a machine-learning model
to implement telematics-based
ADAS. 3) Predict driving safety
using historical data, sensor
fusion, and privacy policy.
AI can help forecast
future EV performance
metrics. Cloud-based
digital twin technology
reduces storage on
mobile systems like
vehicles by decentralizing
data storage.
2021 [18]
6
The AutomationML is a
well-defined modeling tool
and is widely adopted by
companies. It allows
reusability and
interoperability between
different applications and
languages.
A concept of enabling and
supporting CaaS by Digital
Twin was proposed and a
use case has been implemented
to demonstrate how it can be
applied in a real scenario.
Case study illustrates how Digital
Twin supports Car-as-a-Service.
All entities have virtual
representations with middleware.
This system tracks automobile,
user, and user-car data to improve
future user experiences.
Before implementing
this strategy, consider
safety and security.
Blockchain-based
technologies may
help with this.
Wearables can help
build a reactive app.
2021 [19]
7
To handle delays, Petri nets
are used to simulate
development tasks and
dependencies. The model is
connected to and interacts
with the physical system.
Alternative strategies to
overcome delays are
researched, and the ideal
option is computed, tested,
and applied.
The Technical University of
Crete Eco Racing team
employed DTs to design,
manufacture, and assemble
a single-seat urban prototype
vehicle (TUCER). DT is not
a model of the vehicle but is
used to monitor and organize
development tasks.
Real-time interaction between a
physical system and its digital
shadow allows for faster, more
accurate assessments. In case
of job delays, the two systems’
interface calculates and
implements actions to contain
them, computes new time plans,
and tells the user optimistic,
feasible, and pessimistic
possibilities.
Optimizing communication
and interaction between
physical and digital
subsystems, adding
cognition to the Digital
twin, and using enhanced
search methods for
circumstances where no
practical answer can be
determined.
2021 [20]
8
Vehicle to Cloud-based
advanced driver
assistance systems.
A digital twin framework is
proposed for connected
vehicles, which consists of a
physical layer and cyber layer
with various modules. As a
paradigm of this framework,
an advisory speed-based
ADAS is presented using
V2C communication
Three-passenger automobiles
are used to demonstrate the
digital twin framework’s
usefulness in real-world traffic.
The digital twin can improve
transportation mobility and
environmental sustainability
with tolerable communication
delays and packet losses.
This study will evaluate
the digital twin concept
in mixed traffic, where
not all vehicles have
V2C connections. New
service modules must
be built and deployed in
this digital twin topology
to use V2C connectivity.
2020 [21]
9
To create an unsupervised
prognosis and control platform
tailored to electric propulsion
drive systems (EPDS)
performance estimation.
Construct the DT of the energy
system by creating physical
models of its constituent parts
(motors, generator, gearboxes,
bearings, etc.) and their
corresponding reduced models.
Modeling can yield physical
device models (ex. MATLAB).
Physical models are reduced
in order. Parallel, real-time
reduced component models
can DT electric propulsion
drive systems.
Digital twins can act as
virtual sensors or contain
virtual sensors. Combining
real and virtual sensor data
with machine learning can
diagnose electric energy
system devices.
2019 [22]
5

automatic controller or human driver of linked cars to make
cooperative or intelligent motions, enhancing safety, mobility,
and sustainability. The digital twin’s cyber domain computes
this two-layer system. Physical items and processes have
important digital copies. Physical world data is cleaned, inte-
grated, and time-synchronized. Pre-processed data can be kept
in the database for digital traceability or sent to the data mining
& knowledge discovery module for machine learning. Data
mining and knowledge discovery create the physical world
model. Vehicle, driving, and traffic simulators can be modeled.
Data refreshes cyberworld knowledge. Modeling/simulation
aids prediction and decision-making. For system performance,
physical actuators think.
V. COMPARISON ANALYSIS
The findings in Table 1 show the different methodologies
have been mentioned on those selected papers. The third
column defines the role of the DT technologies for the
corresponding methods and their use cases. Then the fourth
column includes the evaluation of those methods. The fifth
column finally includes the future aspects of those pieces of
literary works.
VI. DISCUSSION
A. Application Challenges and Limitations
The difficulties that may develop, according to the authors,
may vary depending on the scope and integration complexity
of the application. Based on the research that was analyzed,
there were five major problems with DT technology implemen-
tation that were found to be universal across all fields. These
problems adequately wrap up the investigation and answer the
supplementary question. SQ1 and contribute to answering the
primary research question as it stands now:
1) Data-related concerns (trust, privacy, cyber security,
convergence and governance, acquisition and large-scale
analysis): If a behavior cannot be reduced to a set of numbers,
it becomes far more challenging for designers to replicate
it. Examples include ecological stability [32], socioeconomic
inequality [33], and political instability [32]. These societal
and environmental innovations will focus on preliminary SRL
stages, where the potential influence on selected stakeholders,
the larger society, and the environment is more understood. In
addition, this difficulty is associated with Table 1’s levels 3 and
4, where the complexity of DT implementations is exacerbated
by the need to enrich models in real time and with bidirectional
flow of information.
2) A deficiency in DT implementation standards, guidelines,
and regulations: Lack of standards and acknowledged interop-
erability, particularly in the manufacturing industry, are cited
as reasons for the current restrictions on DT implementations
by the authors of [34]. Adopting a widespread, concrete
understanding of DTs and their importance requires articles
that explain the benefits, define ideas and structures of DTs,
and review the state of the art of the technology. In addition,
researchers can influence lower levels of the TRL by focusing
on this specific topic through surveys and literature reviews,
thereby increasing the dissemination of fundamental principles
and concepts.
3) High implementation costs owing to more sensors and
computing power: DT implementations are costly, therefore
their use is constrained by the availability of such resources,
which is often lacking in impoverished nations [35]. Achieving
level 3 on Table 1’s maturity spectrum is difficult because
of the rise in required sensors and the resulting increase in
complexity in data connectivity and processing (where the
digital model needs to be enriched with real-time information).
Practitioners are hampered by this difficulty in their pursuit of
greater TRLs, where pilot systems are proven and DTs are
integrated into commercial designs and widespread deploy-
ments.
4) AI and big data for long-term and large-scale data anal-
ysis: Big data algorithms and internet of things technology
are powerful allies that can provide significant support to
successful implementations of DT [36]. This is because DT
systems generate and analyze a significant amount of data,
and these two technologies work hand in hand. In addition
to this, the information that is flowing from the many levels
of indicator systems creates a problem in the process of
defining uniform rules and standards. This challenge aims to
effectively target levels 4 and 5 of the maturity spectrum,
and if it is successful, it may make it possible to enable a
bidirectional flow of information, control of the physical world
from the digital model, and even autonomous operations and
asset maintenance.
5) Challenges associated with communication networks:
The development of superior communication standards like
5G is essential. In [37], the authors discuss the importance of
enabling real-time data connectivity and operational efficiency
for the DT, as well as other benefits of using 5G technology
in smart cities, such as the ability to connect many more
sensors and devices at high speeds with ubiquitous connectiv-
ity, improved reliability and redundancy, and ultra-low power
consumption.
VII. CONCLUSION
The digital twin technology opens up new opportunities for
the development of sustainable electric vehicle technologies
in terms of both cost-effectiveness and efficiency, beginning
with the design phase and continuing through the operation.
Because its sister technologies, such as the internet of things,
decentralized wireless networking, and artificial intelligence,
were not as far along in their development at the time as digital
twin technology was being conceived, the car industry has
only just begun to implement it. Nevertheless, a prime period
for the development of digital twin technologies and other
smart development approaches has now begun in the field of
science, which has now entered this era. In addition, the next
decade will mark a significant turning point in human history
as a result of the extraordinary environmental difficulties
that will be faced. As a result, the fundamental objective
of the research community should be to foster sustainable
technology and the enablers of such technologies. With this
6

goal in mind, the purpose of this review is to demonstrate the
implementation of digital twin technology in the automobile
industry, with a focus on smart electric vehicles serving as
the background for the discussion. In this article, the history
of digital twin technology along with its development and
the many stages of its deployment are discussed. This study
focuses on the digital twin technologies that have been adapted
for smart electric vehicle use cases. Some of these technologies
include predictive mobility, autonomous motion control, driver
assistance systems, vehicle as a service system, Petri net
model discussion, and vehicle-to-cloud-based driver assistance
systems.
REFERENCES
[1] S. Liu, L. Li, J. Tang, S. Wu, and J.-L. Gaudiot, “Creating autonomous
vehicle systems,” Synthesis Lectures Comput. Sci., vol. 8, no. 2, pp.
i–216, 2020, doi: 10.2200/S01036ED1V01Y202007CSL012.
[2] M. Blanco, J. Atwood, S. M. Russell, T. Trimble, J. A.
McClafferty, and M. A. Perez, “Automated vehicle crash rate
comparison using naturalistic data,” Virginia Tech Transp. Inst.,
Blacksburg, VA, USA, Tech. Rep., 2016. [Online]. Available:
https://vtechworks.lib.vt.edu/handle/10919/64420.
[3] N. Kalra and S. M. Paddock, “Driving to safety: How many miles of
driving would it take to demonstrate autonomous vehicle reliability?”
Transp. Res. A , Policy Pract., vol. 94, pp. 182–193, Dec. 2016, doi:
10.1016/j.t ra.2016.09.010.
[4] Hu, Z., Lou, S., Xing, Y., Wang, X., Cao, D. and Lv, C., 2022. Review
and Perspectives on Driver Digital Twin and Its Enabling Technologies
for Intelligent Vehicles. IEEE Transactions on Intelligent Vehicles.
[5] M. Grieves, “Digital twin: manufacturing excellence through virtual
factory replication, ”White paper, vol. 1, pp. 1–7, 2014.
[6] E. J. Tuegel, A. R. Ingraffea, T. G. Eason, and S. M. Spottswood,
“Reengineering aircraft structural life prediction using a digital twin,
”International Journal of Aerospace Engineering, vol. 2011, 2011.
[7] E. Glaessgen and D. Stargel, “The digital twin paradigm for future nasa
and us air force vehicles,” in53rd AIAA/ASME/ASCE/AHS/ASC
structures, structural dynamics and materials conference 20th
AIAA/ASME/AHS adaptive structures conference 14th AIAA,
2012, p.1818.
[8] D. CeArley, B. Burke, S. Searle, and M. J. Walker, “Top 10 strategic
technology trends for 2018, ”The Top, vol. 10, pp. 1–246, 2016.
[9] D. Dolgov, S. Thrun, M. Montemerlo, and J. Diebel, “Practical search
techniques in path planning for autonomous driving,” Ann Arbor, vol.
1001, no. 48105, pp. 18–80, 2008.
[10] B. Yu, W. Hu, L. Xu, J. Tang, S. Liu, and Y. Zhu, “Building the comput-
ing system for autonomous micromobility vehicles: Design constraints
and architectural optimizations,” in Proc. 2020.
[11] Bot´
ın-Sanabria, D.M., Mihaita, A.S., Peimbert-Garc´
ıa, R.E., Ram´
ırez-
Moreno, M.A., Ram´
ırez-Mendoza, R.A. and Lozoya-Santos, J.D.J.,
2022. Digital twin technology challenges and applications: A compre-
hensive review. Remote Sensing, 14(6), p.1335.
[12] Quirk, D.; Lanni, J.; Chauhan, N. Digital Twins: Details of Implemen-
tation: Part 2.AHRAE J.2020,62, 20–24.
[13] Campos-Ferreira, A.; Lozoya-Santos, J.J.; Vargas-Mart´
ınez, A.; Men-
doza, R.; Morales-Men´
endez, R. Digital Twin Applications: Areview.
InMemorias del Congreso Nacional de Control Autom´
atico; Asociaci´
on
de M´
exico de Control Autom´
atico: Puebla, Mexico,2019; pp. 606–611.
[14] Kaur, M.J., Mishra, V.P. and Maheshwari, P., 2020. The convergence of
digital twin, IoT, and machine learning: transforming data into action. In
Digital twin technologies and smart cities (pp. 3-17). Springer, Cham.
[15] Bhatti, G., Mohan, H. and Singh, R.R., 2021. Towards the future of smart
electric vehicles: Digital twin technology. Renewable and Sustainable
Energy Reviews, 141, p.110801.
[16] Yu, B., Chen, C., Tang, J., Liu, S. and Gaudiot, J.L., 2022. Autonomous
Vehicles Digital Twin: A Practical Paradigm for Autonomous Driving
System Development. Computer, 55(9), pp.26-34.
[17] Yang, C., Dong, J., Xu, Q., Cai, M., Qin, H., Wang, J. and Li,
K., 2022, January. Multi-vehicle experiment platform: A Digital Twin
Realization Method. In 2022 IEEE/SICE International Symposium on
System Integration (SII) (pp. 705-711). IEEE.
[18] Niaz, A., Shoukat, M.U., Jia, Y., Khan, S., Niaz, F. and Raza, M.U.,
2021, October. Autonomous Driving Test Method Based on Digital
Twin: A Survey. In 2021 International Conference on Computing,
Electronic and Electrical Engineering (ICE Cube) (pp. 1-7). IEEE.
[19] Steinmetz, C., Schroeder, G.N., Rettberg, A., Rodrigues, R.N. and
Pereira, C.E., 2021, February. Enabling and supporting car-as-a-service
by digital twin modeling and deployment. In 2021 Design, Automation
& Test in Europe Conference & Exhibition (DATE) (pp. 428-433). IEEE.
[20] Tsinarakis, G.J., Spanoudakis, P.S., Arabatzis, G., Tsourveloudis, N.C.
and Doitsidis, L., 2020, September. Implementation of a petri-net based
digital twin for the development procedure of an electric vehicle. In 2020
28th Mediterranean Conference on Control and Automation (MED) (pp.
862-867). IEEE.
[21] Wang, Z., Liao, X., Zhao, X., Han, K., Tiwari, P., Barth, M.J. and
Wu, G., 2020, May. A digital twin paradigm: Vehicle-to-cloud based
advanced driver assistance systems. In 2020 IEEE 91st Vehicular Tech-
nology Conference (VTC2020-Spring) (pp. 1-6). IEEE.Localization and
navigation in autonomous driving: Threats and countermeasures. IEEE
Wireless Communications, 26(4), pp.38-45.
[22] Rass˜
olkin, A., Vaimann, T., Kallaste, A. and Kuts, V., 2019, October.
Digital twin for propulsion drive of autonomous electric vehicle. In 2019
IEEE 60th International Scientific Conference on Power and Electrical
Engineering of Riga Technical University (RTUCON) (pp. 1-4). IEEE.
[23] . Pana, S. Severi, M. Raffero, C. Dannheim, and G. Abreu, “The newest
road revolution: Car as a service,” inAmE 2017-Automotive meets
electronics; 8th GMM-Symposium. VDE, 2017, pp. 1–4.
[24] G. Archer and B. Bondorova, “Does sharing cars re-
ally reduce car use, ”Transport and environment. URL:
https://www.transportenvironment.org/sites/te/files/publications/Doessharing-
carsreally-reduce-car-use-June, vol. 202017, 2017.
[25] F. Ferrero, G. Perboli, M. Rosano, and A. Vesco, “Car-sharing services:
An annotated review, ”Sustainable Cities and Society, vol. 37, pp.
501–518, 2018.
[26] G. Mattia, R. G. Mugion, and L. Principato, “Shared mobility as a
driver for sustainable consumptions: The intention to re-use free-floating
carsharing, ”Journal of Cleaner Production, vol. 237, p. 117404, 2019.
[27] F. Bardhi and G. M. Eckhardt, “Access-based consumption: The case
of car sharing, ”Journal of consumer research, vol. 39, no. 4, pp.
881–898,2012.
[28] M. Q. Khan and S. Lee, “A comprehensive survey of driving monitoring
and assistance systems, ”Sensors, vol. 19, no. 11, p. 2574, 2019.
[29] T. Murata, Petri Nets: Properties, analysis, and applications, Proc. IEEE,
vol. 77, no. 4, pp. 541–580, Apr. 1989.
[30] R. David and H. Alla, Petri Nets & Grafcet – Tools for Modeling
Discrete Event Systems, Prentice Hall, 1992.
[31] J. Wang, Timed Petri Nets: Theory and Application. Norwell, MA:
Kluwer, 1998.
[32] Hirota M, Baba T, Zheng X, Nii K, Ohashi S, Ariyoshi T, Fujikawa H.
Development of new AC/DC converter for PHEVs and EVs. 2011. p.
68–72.
[33] M. Matthew, D. L. O., Castulo, G. Herbert, B. Andrea. Controller -
embeddable probabilistic real-time digital twins for power electronic
converter diagnostics. IEEE Trans Power Electron 2020. PP. 1-1.
10.1109/ TPEL.2020.2971775.
[34] S. Shady. Power electronic converter topologies used in electric vehicles.
2016.
[35] Hirota M, Baba T, Zheng X, Nii K, Ohashi S, Ariyoshi T, Fujikawa H.
Development of new AC/DC converter for PHEVs and EVs. 2011. p.
68–72.
[36] K. Sergiy, T. Galyna, O. Vyacheslav. Genetic algorithms as an optimiza-
tion approach for managing electric vehicles charging in the smart grid.
CEUR Workshop Proceed 2020;2608.
[37] L. Yongkang, W. Ziran, H. Kyungtae, S. Zhenyu, T. Prashant, Hansen
John. Sensor fusion of camera and cloud digital twin information for
intelligent vehicles. 2020.
[38] Kuˇ
si´
c, K., Schumann, R. and Ivanjko, E., 2023. A digital twin in
transportation: Real-time synergy of traffic data streams and simulation
for virtualizing motorway dynamics. Advanced Engineering Informatics,
55, p.101858.
[39] Wang, Y., Su, Z., Guo, S., Dai, M., Luan, T.H. and Liu, Y., 2023. A
Survey on Digital Twins: Architecture, Enabling Technologies, Security
and Privacy, and Future Prospects. IEEE Internet of Things Journal.
7