Available via license: CC BY 4.0
Content may be subject to copyright.
Received XX Month, XXXX; revised XX Month, XXXX; accepted XX Month, XXXX; Date of publication XX Month, XXXX; date of
current version XX Month, XXXX.
Digital Object Identifier 10.1109/OJITS.2022.1234567
Sustainable Intelligent Transportation
Systems via Digital Twins:
A Contextualized Survey
VICTOR M. G. MARTINEZ∗, DIVANILSON R. CAMPELO†, MEMBER, IEEE,
AND MOISES R. N. RIBEIRO∗
1Department of Electrical Engineering, Universidade Federal do Esp´
ırito Santo, Vit´
oria 29075-910, Brazil
2Centro de Inform´
atica, Universidade Federal de Pernambuco, Recife 50740-560, Brazil
CORRESPONDING AUTHOR: V´
ıctor M. G. Mart´
ınez (e-mail: victor.martinez@edu.ufes.br).
This work was supported in part by the Conselho Nacional de Desenvolvimento Cient´
ıfico e Tecnol´
ogico - Brazil (CNPq) under Grant
159528/2019-4, in part by the Fundac¸ ˜
ao de Amparo `
a Pesquisa e Inovac¸˜
ao do Esp´
ırito Santo (FAPES) under Grants 515/2021, 284/2021,
and 026/2022, in part by the Fundac¸ ˜
ao de Amparo `
a Pesquisa do Estado de S˜
ao Paulo (FAPESP) under Grants 20/05182-3 and
18/23097-3, and in part by the Coordenac¸ ˜
ao de Aperfeic¸oamento de Pessoal de N´
ıvel Superior - Brazil (CAPES) - Finance Code 001. The
Article Processing Charge for the publication of this research was covered by CAPES (id ROR: 00x0ma614). For open access purposes, the
authors have assigned the Creative Commons CC BY license to any accepted article version.
ABSTRACT Intelligent transportation systems (ITS) have been attracting the attention of industry and
academia alike for addressing issues raised by the 2030 agenda for sustainable development goals (SDG)
approved by the United Nations. However, the diversity and dynamics of present-day transportation
scenarios are already very complex, turning the management of ITS into a virtually impossible task
for conventional traffic control centers. Recently, the digital twin (DT) paradigm has been presented as
a modern architectural concept to tackle complex problems, such as the ones faced by ITS. This survey
aims to provide a piece-wise approach to introducing DTs into sustainable ITS by addressing the following
cornerstone aspects: i) Why should one consider DTs in ITS applications? ii) What can DTs represent
from ITS’ new physical environments? And iii) How can one use DTs to address ITS SDG related to
efficiency, safety, and ecology? Our methodological approach for surveying the literature addresses these
questions by categorizing contributions and discriminating their ITS elements and agents against the SDG
they addressed. Thus, this survey provides an in-depth and contextualized overview of the challenges
when approaching ITS through DTs, including scenarios involving autonomous and connected vehicles,
ITS infrastructure, and traffic agents’ behavior. Moreover, we propose a functional reference framework for
developing DTs of ITS. Finally, we also offer research challenges regarding standardization, connectivity
infrastructure, security and privacy aspects, and business management for properly developing DTs for
sustainable ITS.
INDEX TERMS Digital twins, intelligent transportation systems, connected vehicles, sustainability.
I. INTRODUCTION
THE transportation sector is among the most important
in modern societies, attracting the attention of industry
and academia to continue its constant evolution alongside
government initiatives. In a simplified way, we can define
transport systems in three fundamental components: the
agents that intervene in the system, the roads through which
the agents move, generally addressed as the road network,
and the operation of the system, which defines the behavior
of traffic and the control systems to manage it [1]. The
2030 Agenda for Sustainable Development, approved by the
United Nations (UN), defined several sustainable develop-
ment goals (SDGs) that include transport as a key enabling
element [2]. Sustainable transport is related to the care of
life and the well-being of people through the road safety
approach (Safety - SDG 3.6), the reduction of polluting
emissions (Ecology - SDG 3.9), the creation and use of
sustainable infrastructures (Efficiency - SDG 9.1), and the
conception of sustainable transport systems that allow access
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
VOLUME , 1
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
to transport in a safe, affordable way for all, improving road
safety (Efficiency, Safety - SDG 11.2).
The smart city paradigm proposes a complex system that,
using basic infrastructure and intelligent solutions, will allow
a highly efficient functioning of humanistic society in several
verticals [3]. Specifically, in the field of transportation,
ITS allows for achieving high efficiency and reliability
in transportation systems through the interaction between
cyber systems and physical transport systems in the context
of smart cities. With ITS, traditional transport systems’
performance is improving through new Internet of Things
(IoT) solutions. They collect data and use it in transportation
system modeling and data analysis for real-time control,
optimization, verification, and validation [4]. However, ITS
is a critical and distributed infrastructure system, making
its management very complex in today’s cities. Therefore,
planning future ITS as a service-oriented architecture goes
well beyond IoT deployment. Scalability, flexibility, and
security are challenging issues for future ITS [5]. In addition,
human factors will significantly impact interaction primitives
and system outcomes [6].
A. MOTIVATION
In recent years, DTs have gained popularity among academia
and industry as a tool for dealing with complex systems.
DTs represent a physical asset in a digitized representation,
which mutually communicates, promotes, and co-evolves
them through bidirectional interactions [7]. Through various
digitization technologies, entities, behaviors, and relation-
ships in the physical world are holistically digitized to create
high-fidelity virtual models. Virtual models formulate real-
time parameters, conditions, and dynamics using real-world
data. This results in a more representative reflection of the
corresponding physical entities, integrating big data analy-
sis, artificial intelligence (AI), and machine learning (ML)
techniques [8]. A bidirectional, reliable, and low latency
communication channel allows the interaction of physical
entities and processes with its DT, exploring the advantages
of cloud computing and softwarization [9]. Although DTs
are being applied to different sectors and activities, the lack
of standard models for physical and virtual entities, data,
connectivity, and standardized architecture for DT makes
their adoption and implementation slow, and their repro-
ducibility and reusability become almost impossible beyond
the borders of a specific solution.
Different standardization bodies have already been trying
to organize the understanding and the modeling of DT’s
intricate features. Their efforts focus separately on topics
such as physical and virtual entities, data, connectivity, and
service features [10]. On the other hand, a holistic approach,
including an open-source initiative, is being pushed forward
by the Digital Twin Consortium 1. It emerged in 2020
as a conglomerate encompassing industry, government, and
academia whose objective is accelerating the development,
1https://www.digitaltwinconsortium.org
adoption, interoperability, and security of DTs and enabling
technologies. The Consortium defined the DT as “a vir-
tual representation of real-world entities and processes,
synchronized at a specified frequency and fidelity” that
allows transforming business models through understanding,
optimal decision-making, and effective action, using real-
time and historical data to represent the past and present and
simulate predicted futures. The main functional blocks of a
DT, namely the physical world, the virtual representation,
and the communication channel, are illustrated for ITS in
Fig. 1.
The main motivation for this survey is the fact that
a DT for ITS (DT-ITS) is a technology with significant
consolidation potential to deal with the increasing number of
isolated frameworks in ITS. This may also be key to evolving
ITS to support current and future compliance requirements
towards SDGs. First, DT-ITS may be a way to deal with
ITS, which is a myriad of sensing platforms in terms of
communicating, storage, and processing. Second, standalone
scenario simulation systems can now be seamlessly inte-
grated into those sensing data within a DT-ITS. It is impor-
tant to highlight that simulations are currently widely used to
optimize transportation networks. They are meant to improve
infrastructure management and support informed decision-
making but presently lack updated parameters to correct
model deviations from real-world behaviors. In contrast, DT-
ITS with simulations integrated into real-time monitoring
of transportation networks [11], such as analysis of traffic
patterns, vehicle movements, and environmental conditions,
will better perform predictions in, for example, what-if
scenarios studies.
Using an in-the-loop simulation strategy, that is, involving
the three blocks depicted in Fig. 1, DT-ITS can more accu-
rately predict the impact of different transportation strate-
gies, helping informed decision-making toward improving
sustainability. Transportation planners and policymakers can
test various scenarios and policies, helping to identify the
most sustainable options, such as changes in land use,
public transport improvements, or the introduction of low-
emission zones [12]. Furthermore, DT-ITS can support life-
cycle assessments of transportation projects, evaluating their
environmental impacts (and compliance requirements) from
construction, operation, and even decommissioning opera-
tions [13]. DT may perform continuous multi-physical mod-
els, i.e., those considering features from mechanical, ther-
mal, electromagnetic, pollution diffusion, and even micro-
biological and human behavior realms [14]. Thus, DT-ITS
yields a unique ever-evolving virtual representation, fed
by sensing, able to take into account aging factors and
weather conditions on a particular physical infrastructure, as
for the road intersection in Fig. 1. This way, present and
future legislation frameworks, such as compliance efforts
toward SDGs, could be studied and implemented, and their
effectiveness verified with the least economic and technical
effort. The hope is that broad and complex challenges,
2 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
virtual representation
computational models
simulation learning prediction visualization
communication channel
synchronization & interaction
physical world
assets & processes
ecology
Sustainable Transport Goals
safety
efficiency
FIGURE 1: The DT paradigm: illustration for ITS.
such as global warming, could be faced by using DT-
ITS as a holistic framework. However, publications on DT
are still scattered across particular domains and specific
applications. Furthermore, hype-driven development in this
area is begging for the right research questions to be asked.
Thus, one may properly understand the specificity of a DT-
ITS implementation and its limitations, with a focus on
developing sustainable transportation solutions.
B. Contributions and Organization
Existing surveys addressing the DT field focus on char-
acterizing DTs from modeling perspectives, architecture
proposals, and the categories of services and applications
most used in a field. Readers are referred to surveys that
deal with foundation concepts of DTs [15]–[22]; industry
verticals [23]–[26] and also in Industry 4.0 manufacturing
models [27]; civil engineering [28]–[33]; agriculture [34]–
[36]; energy [37]–[40]; healthcare [41], [42]; and, finally,
to smart cities [43], [44]. Although DTs have been gaining
popularity in transportation system solutions, only a few
surveys in the literature update the state of the art of
DTs for intelligent transportation systems. For instance, a
recent comprehensive “survey of surveys” on milestones
in autonomous driving and intelligent vehicles by L. Chen
et al. [45] could only locate a single survey on DTs in
that context, i.e., [14]. Thus, it is clear that there is a
need for a comprehensive survey that collects and orga-
nizes the trends in using DTs as a transitioning element
toward intelligent transportation solutions. Surveys on DTs
for transport systems have recently been published [?], [46]
[47]. Existing surveys addressing the DT field focus on
characterizing DTs from modeling perspectives, architecture
proposals, and the categories of services and applications
most used in a field. Readers are referred to surveys that
deal with foundation concepts of DTs [15]–[22]; industry
verticals [23]–[26] and also in Industry 4.0 manufacturing
models [27]; civil engineering [28]–[33]; agriculture [34]–
[36]; energy [37]–[40]; healthcare [41], [42]; and, finally, to
smart cities [43], [44].
Although DTs have been gaining popularity in transporta-
tion system solutions, only a few surveys in the literature
have updated the state of the art of DT-ITS. For instance,
a recent comprehensive ”survey of surveys” on milestones
in autonomous driving and intelligent vehicles by L. Chen
et al. [45] could only locate a single survey on DTs in that
context, i.e., [14]. Thus, it is clear that there is a need for a
comprehensive survey that collects and organizes the trends
in using DTs as a transitioning element toward intelligent
transportation solutions. Only very recently surveys narrow-
ing on DT-ITS [46], [47] can be found, but they still fail to
bring a critical discussion on the actual role of DTs in the ITS
context and their (functional and non-functional) require-
ments. Furthermore, they lack a broader view of transport
challenges like those brought by SDGs, such as efficiency,
safety, and ecology. Such latest survey propositions focus on
their applications and ignore real-world implementations and
case studies initiatives that already apply DT-ITS. Finally,
another contribution here is that we managed to translate
challenges (alongside their requirements) and the lessons
learned from this comprehensive survey into a framework
aligned with the International Telecommunication Union
Telecommunication Standardization (ITU-T) recommenda-
tions [48]. This keeps our discussion context within the UN’s
diverse efforts: from sustainable goals to standardization-
compatible DT-ITS framework proposition.
In summary, the following points are the principal contri-
butions of this article:
1) It gradually unveils the importance of DT technology
by framing the recent advances in the UN’s sustain-
ability goals and identifying the early adopters and key
areas of modern ITS that DTs will empower.
2) It presents a reference model that summarizes the
lessons learned into a DT-ITS framework with its func-
tional and non-functional requirements and research
directions that need contributions.
The methodology used is based on research questions
that would outline the course of the survey, as well as the
VOLUME , 3
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
revision protocol to be applied. The survey aims to identify
current trends in adopting DT throughout the vehicular trans-
port ecosystem, identifying architectures and technologies
that make these solutions more efficient than traditional
approaches in the context of smart cities. Having this in
mind, we posed the following research questions (RQ):
•RQ #1: Why should one consider DTs in ITS applica-
tions?
•RQ #2: What can DTs represent from ITS’ new physical
environments?
•RQ #3: How can one use DTs to address SDG regard-
ing efficiency, safety, and ecology in ITS?
We surveyed relevant publications from 2018 in the
Scopus, Web of Science, and IEEE Xplore databases and
complemented them with other works from Google Scholar.
We organized the selected papers and discussed them se-
quentially to provide answers to the RQs above. We consider
ITS’ twinned entities and sustainability goals on key areas
empowered by DTs, providing a piecemeal approach to grad-
ually laying solid foundations and motivations for readers to
contribute with their research efforts toward sustainable DT-
ITS.
The remainder of this work is organized as follows.
Section II discusses adopting DT technology as an alternative
to address known problems in the design, operation, and
management of transportation systems. Section III describes
the research background and status of DTs as enablers of a
new generation of intelligent transportation solutions. Sec-
tion IV introduces a reference framework to accurately and
timely develop DT-driven applications in transportation sys-
tems. Section V discusses challenges that require substantial
research efforts and careful planning to enable sustainable
DT-ITS. Finally, Section VI concludes this survey.
II. PRECURSORS FOR DIGITAL TWINS IN ITS
DT-ITS are defined as the digital representation of the trans-
portation system, including intelligent infrastructure, traffic
participants, traffic behavior, and the surrounding environ-
ment. To achieve a fidelity representation, new advanced
enabling technologies are employed for data collection,
processing, privacy, and security, using accurate and real-
time data collected. A key difference DTs can make for
ITS is that the verisimilitude of their virtual models may
allow for an early and even anticipated discovery of problems
in transport systems. Thus, short-, medium- and long-term
strategies can be outlined to support decision-making and
achieve more robust ITS applications.
Even though DTs have marked a notable presence in
academia, the industry has perceived the adoption of this
technology to be slower, and this begs RQ #1 on the practical
relevance of DTs for current and future ITS demands. This
section presents contexts that may drive early adopters to see
DTs as a graceful evolution pathway for legacy techniques
in ITS instead of a disruption to current practices. Thus, RQ
#1 is here addressed by focusing on i) sensing consolidation
and scenario simulation; and ii) test systems using in-the-
loop approaches.
A. SENSING CONSOLIDATION AND SCENARIO
SIMULATION
ITS is now an information-rich scenario with extensive sens-
ing coverage of roads and connected vehicles so that human
drivers can be better informed about dangerous traffic, road
conditions, adverse weather conditions, and traffic conges-
tion. DTs may become a framework to enhance current
applications to meet UN sustainability goals. For instance,
real-time and detailed road and traffic information allows
road authorities to exercise more precise traffic control, such
as lane-level traffic control and ramp merge, to improve
safety and traffic efficiency. Using IoT sensing devices,
combined with digital maps and road-building information
models [49], DT systems can be used to consolidate road
traffic systems and road health and asset-monitoring systems
[50].
Nevertheless, the analysis of all the possible situations in
a transport scenario involving several agents, vehicles, and
non-vehicles is diverse, making their evaluation very difficult
from the current model of traffic control centers. Presently,
simulators are used as flexible and efficient tools for ITS
to predict scenarios, but it is well known that they need
proper traffic representation. The strong coupling between
the vehicle, the environment, and other agents is oversim-
plified. Moreover, the amount of computational resources
demanded for modest scenarios is presently a bottleneck for
ITS. In this context, a method to enhance the credibility of
simulation-based testing for automated driving was proposed
by Stadler et al. [51]. The method performs a real test drive
with automated driving functions. Then, the scenarios are
identified from the real driving data and re-simulated, in-
cluding the static environment and all traffic participants. In
this way, a qualitative and quantitative comparison could be
carried out between virtual scenarios and different simulation
setups. Besides, Liu et al. [52] proposed a new pipeline to
create artificial scenes and generate virtual datasets based
on parallel vision theory with low modeling time and high-
quality labeling.
DTs have been proposed in the operational design domain
(ODD) analysis of such systems for security validation in
autonomous systems. Sun et al. [53] proposed an architecture
design to improve the ODD of autonomous vehicle systems.
In this way, it is possible to capture the operational con-
straints of the generic components of the road environment,
such as road type, traffic volume, and weather conditions.
Determining which ODD to use for an automatic driving
system function can be compared to finding the driving
environment condition boundaries that satisfy particular eval-
uation criteria based on potential scenarios, allowing more
realistic environment models to be created. To represent
structural, physical, and behavioral information in a virtual
4 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
traffic-in-the-loop
vehicle-in-the-loop
human-in-the-loop
model-in-the-loop
FIGURE 2: Current in-the-loop ITS’ test systems and simulations involving agents, communications, and intelligent
infrastructures
environment, Yu et al. [54] developed a game engine-based
DT system that provides physics and graphics engines for
3D modeling, image rendering, and physical simulation.
The physics-based models derived were well-behaved
to solve physical processes and systems for a long time.
However, many complex systems escape such quantitative
analytical descriptions or the correct selection of input
variables [55]. The high computational demand required to
solve such models makes them unsuitable in most cases
for real-time ITS applications. Thanks to a large amount of
data and the rapid development of AI/ML solutions, data-
driven models are presented as an excellent complement
to physics-based models, mainly in supporting optimization
and simulation tasks. This is why a balanced perspective
of both approaches is needed for complex transportation
systems. Combining both represents the most successful way
to construct DT-based models for ITS. Compared with con-
ventional transportation simulation, DTs have the potential to
improve precision, ease of implementation, and digitization
of procedures [56].
The predictive power of the so-called data-driven models,
like AI/ML solutions, depends on historical data for training.
Thus, it is severely limited by the number of (rare) events
that have occurred before. For sustainability goals related to
safety, predicting rare events is crucial. Some agent-based
modeling approaches best provide plausible scenarios for
situations that have not happened before. However, they still
need to be more effective in their predictive power [57].
Recent trends in symbiotic simulation studies emphasize
its combination with machine learning. Despite its success
and usefulness, very few works focus on applying a hybrid
system of this type in microscopic traffic simulation. The
application of ML models in microscopic traffic simulation
is limited to predictive analytics or offline simulation-based
prescriptive analytics. Therefore, DTs enabled by data-driven
models help to dynamically update the parameters of the
deep learning model for real-time traffic simulation [58].
A pay-as-you-go approach can be adopted economically
from legacy systems to DT-oriented ITS. DTs will become
increasingly interoperable. Early DT designs focused on
individual domains. New DT interoperability standards will
facilitate the composition of larger-scale DT assemblies from
a library of designs. Standards will accelerate efforts to
reuse DT components across multiple designs in ITS. More-
over, new wireless communication technologies, especially
vehicle-to-everything (V2X) communications, will play a
decisive role in DT interoperability. V2X is already crucial in
the modern ITS test system, so we argue this is an important
precursor for DT adoption.
B. TEST SYSTEMS USING IN-THE-LOOP APPROACHES
To support the V2X road test, Han et al. [59] designed
a live transmission system on the road, which simulates
the information of non-V2X-equipped road vehicles through
the V2X array node. A road sensing system was used to
collect information from the vehicle data to be used in
procedures for reconstructing scenarios based on V2X and
extracting road scenarios. Playback of critical scenes could
be considered a type of DT test; the reliability of the test
results largely depends on the validity of the test scenario.
A critical scenario extraction method is proposed to meet the
test requirement. Like the live broadcasting function, for the
playback function, the test scenario is expressed by a C-V2X
message and disseminated via node array to simulate V2X
application scenarios.
The DT-oriented test system can consolidate simulation
efforts and thus has been increasingly used to meet vehicle
navigation and security requirements [60], in many cases
using simulation loop schemes such as model-in-the-loop
(MIL), scenario-in-the-loop (ScIL) [61], [62], vehicle-in-the-
loop (VIL) [63], and human-in-the-loop (HITL) testing [64],
VOLUME , 5
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
[65], as shown in Fig. 2. Note that in-the-loop approaches
involve the physical world, assets and processes, and com-
putational models while using communication channels for
interaction. In other words, it can be seen as a prototype that
fits the DT in Fig. 1.
Many simulation platforms use game engines to render
a fairly realistic 3D scene to enhance the realism of the
simulation scene. However, rendering this way is expensive
and resource-intensive, and the data obtained from some
types of virtual sensors is very different from the real scene.
This problem was addressed in [66], which proposed an
injection simulation framework capable of customizing roads
and traffic scenarios to achieve SIL and vehicle-in-the-loop
(VIL) tests for different functional modules. The scenario
generation module is based on the real road’s high-precision
semantic map (OSM map) and the simulator SUMO to define
the test scenario. The framework could inject data into the
fusion and perception layers by providing accurate simulated
traffic information and LIDAR data generated in real time.
Besides, Bal´
azs et al. [62] presented a novel simulation
concept for ITS called scenario-in-the-loop (SciL) testing.
The SciL concept is the extended version of the traffic-in-
the-loop (TiL) simulation method, created to test autonomous
driving functions in critical collision situations using real
targets.
A DT-based automated driving test method was proposed
by Shoukat et al. [67]. Vehicles collect and release driving
information through V2X communication, execute the data
fusion, and upload the information to the simulation plat-
form. With a similar approach, a mixed reality simulation
environment was introduced by Szalai et al. [68] that inte-
grates a real test vehicle into a virtual environment. Thus,
the behavior of a real vehicle connected to the simulation
can be tested in real time with a VIL approach. With VIL,
the real movement of the test vehicle allows testing the
vehicular functions at the decision and movement planning
level, even at low costs for automotive manufacturers and
researchers. Shuguang et al. [69] also proposed a novel
VIL verification method based on vehicle-road-cloud col-
laboration. In the solution, the autonomous driving obstacle
avoidance algorithm is verified in the highway scene using a
mixed scene. Another motivation for using DT to transform
future ITS applications is vehicular and pedestrian mobility.
Understanding the global behavior of urban transportation
allows the construction of projections for decision-making
stakeholders, i.e., traffic management entities and public
policy formulators. The classic features of human mobility,
such as daily variation, have already been addressed by DTs
to mitigate restrictions imposed on both mobility and agent
rationality when classic mobility models are generated [70].
By revisiting RQ #1 (“Why should one consider DTs
in ITS applications?”), we can conclude that DTs are a
pathway for fragmented assets like sensing, computation,
and communication so that simulation and testing techniques
can be appropriately consolidated into a single framework
for designing and operating future transportation systems.
The rise of domain-specific languages (DSL) for modeling
virtual entities has significantly promoted the adoption of
DTs, which are a powerful tool for the agile development
and evaluation of ITS applications. This evolution in design
methodology has made it possible to analyze all possible
scenarios swiftly and manage transport systems more effi-
ciently. This way, DTs can help meet the most critical goals
for raising sustainable transport systems.
III. KEY AREAS EMPOWERED BY DIGITAL TWINS IN ITS
This section introduces a comprehensive survey on DT-ITS
to answer RQ #2, considering now the main actors of the
future transport ecosystems. From the selected papers, the
ITS new physical environments benefiting from DTs are
as follows: Connected and autonomous vehicles, intelligent
transport infrastructures, ITS agent behavior, and the Internet
of vehicles. The context for these key application areas is
illustrated in Fig. 3, which expands the perspective and
concepts brought in Fig. 1 for the physical world. To
address what can be represented by DTs in such areas the
main contribution of the surveyed papers are presented in
Tables 1, 2, 3, and 4 also supported by the comparative
charts highlighting twined elements and the sustainable goals
spanned in Figures 4, 5, and 6, respectively.
A. CONNECTED AND AUTONOMOUS VEHICLES
New vehicle concepts have begun to emerge in parallel
with advancements in other branches of technology, making
connected cars a reality today in hybrid scenarios, such as
the one illustrated in Fig. 3a, which features connected and
autonomous vehicles. Developing efficient decision support
systems, driven by customer-centric approaches empowered
by DT, plays an essential role in decision-making within
automotive engineering [71]. Specifically, implementing ad-
vanced driver assistance systems (ADAS) significantly im-
proves safety for drivers and the entire traffic system. Cur-
rently, the most common assistance systems only suggest
optimal actions for drivers. However, solutions that can take
control of the vehicle to prevent accidents or minimize
their consequences are becoming increasingly prevalent. DT-
enabled solutions can be applied to ADAS systems to achieve
more efficient levels of assistance for drivers [64].
Wang et al. [72] proposed a framework that uses DT
to provide drivers with recommended speed information
through an onboard interface, allowing more efficient vehicle
control. A DT of the traffic scenario, including the physical
road infrastructure and the transport agents, was developed
in the cloud. Models of a ramp merging use case roll
interacted with the physical world through vehicle-to-cloud
communication. This work was extended by Liao et al. [49],
adding the modeling of the human behavior of the driver.
With the support of this DT, a reduction in average speed
variation was obtained, which shows that the cooperative
fusion approach of the sensor system used is safer than in
6 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
connected
vehicles
connected
vehicles
autonomous
vehicles
(a) Connected and autonomous vehicles.
intelligent
infrastructure
sensing and
actuation
(b) Intelligent transport infrastructures.
agents behavior
conflict resolution
routes update
(c) ITS’ agents behavior.
V2I
communication
IoV Management
Vehicular
network
(d) Internet of Vehicles.
FIGURE 3: Key application areas empowered by DT in ITS.
the reference scenarios without driver assistance. In addition,
reducing polluting emissions and fuel consumption was ver-
ified, complying with lines of action for creating sustainable
transportation.
In ADAS systems, adaptive cruise control (ACC) adjusts
the longitudinal speed of vehicles to maintain a safe dis-
tance from the vehicle in front, increasing safety, enhancing
driving comfort, and reducing fuel consumption. However,
the functionalities of ACC limit the following drivers’
styles, resulting in a lack of confidence in ACC systems
and limiting their adoption. A method supported by DTs,
which learns from drivers’ natural vehicle tracking behavior,
was proposed by Wang et al. [73]. The model generates
a vehicle acceleration profile that adapts to the driver’s
preferences, resulting in a personalized ACC system that
contributes to collision avoidance. HITL experiments on a
driving simulator validated the solution. Techniques that use
DTs to validate HITL teleoperation in transportation systems
were proposed by Kuru [74]. A case study of personalized
adaptive cruise control is also used by Wang et al. [75]
to evaluate the potentialities of the game engines platform
in creating complete models of connected and autonomous
vehicles (CAV).
Although ADAS are today a reality in solutions from
several vehicle manufacturers, these systems suffer from
unnecessary warnings or propose strange actions for drivers,
especially in complex traffic scenarios. Moreover, false pos-
itives cause driver distraction and confusion, posing addi-
tional safety concerns and undermining such systems’ credi-
bility. Mindful of this problem, Tang and Jiang [76] proposed
a solution that infers the driver’s perception through gaze
tracking to improve the driving assistance system. Based
on historical observations, the proposal mimics the driver’s
prediction of unobservable agents in the driving environment,
warning the driver only if an agent represents an unperceived
imminent collision threat.
Only real-time information from the target vehicle can be
perceived when ADAS solutions are based solely on object
detection sensors. The last can result in historical data being
unable to predict vehicle behavior due to the short detection
time horizon. To improve the sensor fusion process used
in ADAS systems, a new sensor fusion methodology that
integrates images obtained by cameras and the knowledge
learned by DTs executed from the cloud was introduced
by Liu et al. [77]. The authors predicted the lane change
behavior of vehicles around the ego vehicle to validate the
solution. HITL simulations occur in an intelligent vehicle
simulation environment based on a game engine. In addition,
an information visualization method based on augmented
reality (AR) was adopted as part of the DT to display the
predicted information found on the cloud. The work was
extended in [78], where a multi-layer perceptron algorithm is
proposed with modified lane change prediction approaches.
Simulation results revealed that the proposed model could
significantly improve highway driving performance regard-
ing safety, comfort, and ecological sustainability.
In autonomous vehicle scenarios, DT approaches regard-
ing driving strategies in dynamic environments have been
studied to guarantee the safety of the entire transport system.
Driving is a social activity that involves endless interactions
with other agents on the road, so locating these agents
and predicting their possible future actions in the envi-
VOLUME , 7
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
ronment is essential for the safety of all agents. A crash
mitigation algorithm that directly incorporates a generalized
crash severity index model for vehicle-to-vehicle collisions
of multiple impact patterns was proposed by Li et al. [79].
The algorithm tries to adapt the vehicle position before the
collision to minimize the impact’s severity. The algorithm
was tested in a real-vehicle intersection crossing scenario
and validated through DT simulations. Wang et al. [80]
designed a road driving safety analysis based on extracting
vehicle movement information using drone images. In this
way, scenes of vehicle movement in the virtual world and
analysis of vehicle driving risks can be built. The results
showed that driving risk analysis based on the DT method
could more accurately monitor various status indicators in
the vehicle movement process.
Yang et al. [81] proposed a DT method for executing
multi-vehicle experiments, using a combination of physical
and virtual vehicles to perform coordination tasks. For this,
a sand table testbed was developed with its DT in the
cloud, where devices enable human-machine interaction, all
visualized in a mixed-reality solution. The study explores the
concept of cloud vehicles, allowing the control of vehicles
from the first-person perspective in the driving simulator.
Another study on the prediction of driving states of vehicles
to enhance the accuracy of traffic safety was conducted by
Lv et al. [82]. In this solution, the vehicle simulator and
the virtual environment system are built based on vehicle
dynamics through virtual reality technology. A DT of the
vehicle was built based on various sensors and a Gaussian
process algorithm.
DT solutions have widely addressed the safety of transport
agents. A method of predicting and preventing traffic acci-
dents based on DTs and artificial intelligence to guarantee
the safety of drivers and pedestrians was proposed by Lv et
al. [83]. A neural network-based object tracking algorithm
is applied to DTs for video analysis in traffic accident
detection. The solution used computer-aided design (CAD)
software to model transportation drawings on 3D models.
Intelligent security systems have also been proposed by Lu et
al. [84] to create virtual entities that act as co-pilot systems.
Solutions of this type are designed to make the real-time
evaluation, response, and recording of failures possible. With
digital co-pilot systems, it is possible to provide analysis
and feedback and record incorrect decisions, which helps in
the self-adaptation of the system. Other design and imple-
mentation methodologies of DTs for safety validation and
automotive safety have been proposed in [85], [86].
Cooperative driving at non-signalized intersections has
been a popular topic in intelligent transportation systems
research. In these new intersections, the driving strategies
adopted by AV represent a crucial point for the safety
of the agents interacting in the conflict zone of the in-
tersection. Wang et al. [89] proposed a DT architecture
for AV and human-powered connected vehicles where the
software modules and their algorithms are developed in the
Transport SDGs
Traffic
Environment
Vehicle
Driver
Efficiency Safety Ecology
[91]
[72] [74]
[73] [83]
[84] [92]
[14]
Twinned
Entity
[49]
[76]
[78]
[78]
[77] [75]
[79] [82]
[90]
[85] [86]
[80] [89]
[88]
[87]
reference {twinned entity
reference {twinned entity transport SDG}
transport SDG}
For references address more than a Transport SDG
or more than a twinned entity:
FIGURE 4: Main research on DT for CAV classified by the
transport SDG they address.
digital world. The vehicle’s DTs are displayed to the driver
using AR, enabling them to engage in cooperative driving
behavior with other CAVs at non-signalized intersections.
The simulations were performed using game engines, where
HITL simulations are conducted with the guidance provided
by AR. With the development of a DT-enabled architecture to
facilitate collaborative and distributed autonomous driving,
Hui et al. [90] introduced the concept of collaboration-as-a-
service, designing a DT for each AV. With this architecture,
a collaboration mechanism based on auction games was
developed to decide the lead DT and the tail DT in each
driving group. In addition, DTs can replace AVs to make
collaborative driving decisions in virtual networks in ad-
vance, where frequent information exchanges between AVs
in physical networks can be avoided. Kumar et al. [87] also
deduced the driver intention from collecting various data and
a virtual vehicle model in the cloud.
Accurately predicting a driver’s behavior remains a chal-
lenging problem in road safety [93]. Cheng et al. [88]
proposed a framework to address this problem, in which
driver behavioral twins are shared between connected cars
to predict the potential actions of neighboring vehicles, thus
improving driving safety. The behavior of a vehicle is deter-
8 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
TABLE 1: Summary of main research on DT-empowered CAV.
Ref. Year Main contribution Weakness and limitations
[87] 2018 An AI-based solution to precisely and perfectly measure the
traffic situation in real-time and the driver intention to improve
traffic efficiency.
Does not adequately address critical challenges associated with
reliability, security, and privacy.
[88] 2018 A framework in which behavioral models of drivers are shared
among connected cars to predict potential future actions of
neighboring vehicles.
Lack of discussion on practical considerations related to scala-
bility, infrastructure, and connectivity.
[85] 2019 A solution to predict driving patterns, typical routes, and driver
habits, designed to augment customization of smart cars and user
driving experience.
Overlooks the significant resources and infrastructure required
for real-time data processing in large-scale autonomous vehicle
systems.
[72] 2020 Recommendation of speed information through an onboard in-
terface for road merge scenarios.
The key concern is the insufficient focus on dependability,
security, and privacy.
[76] 2020 A driving assist system that reduces unnecessary warnings by
taking into account the driver’s perception of the driving envi-
ronment.
underexplored governance, business, and human factors leaving
critical gaps in its proposed DT and raising significant questions
about the system’s feasibility and trustworthiness.
[77] 2020 A sensor fusion methodology, integrating camera image and DT
knowledge from the cloud to help intelligent vehicle decisions.
It lacks sufficient consideration of potential vulnerabilities in
transmitting and processing sensitive data via the cloud.
[74] 2021 Human-on-the-loop haptic teleoperation for human control of
remote vehicles collaboration.
Risks associated with security and privacy breaches are not
addressed.
[75] 2021 A DT simulation architecture based on Unity game engine that
integrated analytic external tools.
It fails to address vulnerabilities in data transmission between
Unity and AWS or measures to protect sensitive driver data in
the cloud.
[86] 2021 A framework for vehicular DTs that facilitate the data collection,
data processing, and analytic phases.
Does not adequately address scalability, infrastructure, and con-
nectivity, highlighting the need for a more comprehensive dis-
cussion on practical implementation hurdles.
[49] 2022 Recommendation of speed information, an extension from [72]. Focus on a specific set of assumptions and a custom-built
system without clear integration with standardized frameworks
or broadly applicable evaluation metrics.
[73] 2022 Learn from drivers’ naturalistic car-following behavior and out-
puts an acceleration profile that suits the driver’s preference.
It doesn’t delve into potential vulnerabilities within the data
acquisition, communication channels, and the algorithm itself.
[78] 2022 Extension from [77]. Lack of in-depth analysis of the scalability, infrastructure, and
connectivity challenges required for the practical implementation
of the proposed system.
[79] 2022 A motion planning scheme to handle the crash mitigation in
scenarios with unavoidable collision.
Does not discuss the communication-specific protocols to address
the dependence on accurate and timely information.
[80] 2022 Evaluation of driving risks according to the stability and trajec-
tory deviation of the vehicles.
Faces scalability, security, and governance limitations, with con-
cerns about reliance on large-scale deployment.
[82] 2022 Moving vehicles tracking and recognition algorithm by video to
predict the vehicle movement integrating VR and DT.
Does not address the crucial aspect of real-time interaction
between the DT model and the physical entity.
[83] 2022 An AI-based traffic accident prevention and prediction system
through data provided by a DT.
Fail to address the critical security and privacy implications
inherent in such a system.
[84] 2022 A co-pilot intelligent decision safety system to better supervise
autonomous driving systems with online cloud learning.
Overlooks practical considerations, such as lifecycle management
and business and cost implications, for clear governance struc-
tures.
[89] 2022 A consensus motion control algorithm to generate the vehicle
speed trajectories in non-signalized intersections.
Does not address connectivity issues such as latency variation,
inaccuracy, and synchronization issues, without specific commu-
nication technologies.
[90] 2022 An efficient collaborative autonomous driving scheme to mini-
mize the cost of completing the driving process.
Insufficient detail regarding its dependability and security in
dynamic vehicular networks.
[14] 2022 A study on driver DTs and its key enabling aspects. Data Governance and Privacy are vaguely addressed.
[91] 2022 A virtual reality enabled by DT for on-road emission monitoring
for developing and evaluating eco-driving strategies.
It overlooks the vulnerabilities of real-time data. Privacy concerns
related to vehicle data are mentioned briefly.
[92] 2023 A shared controller for safe and human-friendly cooperative
driving based on predictive risk assessment.
Does not address standardization, scalability, reliability, security,
privacy, governance, or human factors.
VOLUME , 9
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
mined by its driving context, which includes road conditions,
nearby agents, infrastructure, and even the driver’s state of
mind. However, the driving context perceived by the human
driver may differ from that perceived by the vehicle, making
it even more challenging to predict driver behavior. The
proposal collects historical driving data to build a driver
behavior profile model for each vehicle, which can be used
to predict its future behavior in different driving contexts. A
virtual platform was developed in a game engine platform
to eliminate the effect of unrelated factors.
In this regard, it is essential to build a unified human-
centered intelligent driving system that considers the proac-
tivity and sensitivity of the human driver. The digitization of
the human driver must be able to model habits, personality
traits, and decision patterns, which are crucial characteristics
that define a human driver. Hu et al. [14] introduced the
first driver DT model. The proposal attempts to bridge the
gap between existing automated driving systems and fully
digitized ones and assist in developing a complete cyber-
physical human driving system. With the driver DT model,
it is possible to monitor the physiological state to help
prevent accidents due to distraction or sudden illness and
improve driving safety. Key features of this driver DT include
multimode state fusion, custom modeling, and time variation.
This system provides the autonomous vehicle with improved
personalization, allowing it to drive like a human being.
B. INTELLIGENT TRANSPORT INFRASTRUCTURES
Sustainable urban road planning should strive to meet current
and future traffic-related demands and achieve financial,
environmental, and social benefits, a complex and inter-
disciplinary subject. Intelligent and adaptive transportation
infrastructures, as shown in Fig. 3b, are one of the bases
for achieving truly sustainable transportation solutions. A
toolchain for visualizing detailed road traffic data from mul-
timodal sensors was proposed by Neuschmied et al. [94]. The
approach was based on an audiovisual analysis of the traffic
scenario. It provides a holistic view, including tracking and
counting traffic participants, detecting potentially dangerous
situations, and analysis of sources of noise pollution. A
spatial data management system extracts information with
a geographic information system (GIS), and a web-based
viewer allows interactive visualization in the context of
a high-definition DT of the traffic environment. Likewise,
Jiang [50] proposed a DT-based urban road planning ap-
proach employing multi-criteria decision-making and GIS.
The urban planning framework considers land use, traffic
congestion, driving route selection habits, air quality, and
noise. The approach could obtain functional, economical,
and environmentally friendly urban road planning. Unlike
previous studies, the proposal considers the construction of
new roads and includes an analysis of existing roads. Another
DT system based on GIS technology was proposed by Wang
et al. [95] to provide a new way of surveying and mapping
in the transportation industry and transform the horizontal
reference {twinned entity
Transport SDGs
Traffic
Environment
Transport
Network
Vehicles
Efficiency Safety Ecology
Twinned
Entity
[97] [102] [95] [103]
[106] [99][96] [101]
[104]
[100] [105]
[98]
[50]
[94]
[98] [105]
reference {twinned entity transport SDG}
transport SDG}
For references address more than a Transport SDG:
FIGURE 5: Main research on DT for intelligent transport
infrastructures classified by the transport SDG they address.
and vertical design of highways into computer-aided design.
Also, Guo et al. [96] proposed a 3D digital system based
on roadside sensing of a cooperative vehicle infrastructure
system, allowing the visualization of road traffic in real time.
DT systems for road traffic, road health, and asset moni-
toring can be built using IoT devices combined with digital
maps and road construction information models. Mao et
al. [98] offer a future roadmap for developing IoT systems
to construct DTs of intelligent roads. The CitySim database,
formed from the vehicular trajectories obtained through
videos recorded by drones, was created by Zheng et al.
[106] to facilitate the research and development of security-
based applications. CitySim facilitates research towards DT
applications by providing relevant assets such as 3D base
maps of recording locations and signal times. CitySim’s
trajectories were generated through a five-step procedure
that ensured trajectory accuracy. In addition to base maps,
CitySim provides signal timing data related to signalized
intersections. The calibrated traffic patterns and 3D base
maps are used in a collaborative simulation that integrates
microscopic traffic and driving simulations.
A DT system for traffic flow perception and risk identifi-
cation has also been proposed to improve ITS infrastructure.
A road and traffic trajectory model for highway entrances
and exits was proposed by Liu et al. [99]. The model is
also built based on drone information and computer vision
methods to extract microdata of traffic flow, including driving
trajectories and vehicle speeds. Moreover, Ji et al. [100]
designed the DT of a road network to observe the traffic op-
10 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
TABLE 2: Summary of main research on DT-empowered intelligent transport infrastructures.
Ref. Year Main contribution Weakness and limitation
[95] 2021 An analysis of the modeling process of the road traffic DT
system to verify the feasibility of the practical application using
geographic information.
It overlooks the practical infrastructure, scalability, and connec-
tivity challenges. Dependability, security, and privacy concerns,
such as vulnerabilities in physical systems and data protection,
receive minimal attention.
[96] 2021 A platform for real-time roadside sensing that uses the Robot
Operating System to create 3D virtual road traffic environments.
It doesn’t discuss standardized frameworks, scalability chal-
lenges, and hardware dependencies. It focuses on a specific
implementation.
[97] 2021 A DT platform for the digital twinning of roads. It fails to address reliability, security, and privacy, particularly
in managing inherent vulnerabilities. Do not consider scalability
costs.
[94] 2022 A toolchain for visualizing detailed road traffic data from multi-
modal sensors.
It neglects aspects such as scalability, security, and governance.
[50] 2022 A sustainable urban road planning approach that uses data
from multiple sources in the physical world to assist new road
construction and old road widening.
The lack of a feedback loop, the lack of a plan to update the
DT, and the absence of discussion of the framework’s scalability
are significant limitations. These omissions limit the practical
applicability and long-term usefulness of the proposal.
[98] 2022 An IoT-based system for smart roads to enable the construction
of a DT of the traffic and road system and various applications.
There is a lack of discussion on infrastructure and connectivity
issues, including latency, imprecision, and synchronization prob-
lems, as well as insufficient communication standards.
[99] 2022 A drone sensing scheme with machine vision methods used to
extract traffic flow micro-data, including driving trajectories and
vehicle speeds.
It does not discuss in detail the application or interaction with
real-world scenarios, nor the scalability issue or infrastructure
requirements.
[100] 2022 A model-free method that uses macroscopic road network images
to form an encoding-decoding structure to predict spatiotemporal
congestion caused by accidents.
The paper’s limitation lies in its limited exploration of scalability,
infrastructure, and connectivity.
[101] 2022 Develop a framework for a DT smart freeway. Lacks in-depth analysis of real-world challenges, especially re-
lated to scalability, infrastructure, security, and governance.
[102] 2022 A scheme to integrate third-party controllers to dynamically
generate and calibrate traffic flow in simulation scenarios in
SUMO.
The paper emphasizes the use of real-time data for adjusting the
simulation, but it does not delve into the potential challenges
related to connectivity, latency, scalability, security, and privacy.
[103] 2022 A framework for DT in intelligent intersection modeling to
consolidate real-time monitoring, control, and management of
road intersections.
It provides limited details on addressing challenges related to
standardization, scalability, and connectivity, particularly in im-
plementing a robust architecture, managing multi-tier computing,
and ensuring low-latency communication.
[104] 2023 A novel Blockchain-based smart parking scheme in DT empow-
ered VSNs with privacy protection.
Does not perform stress testing on its blockchain component,
leaving scalability and performance under high-stress conditions
unaddressed.
[105] 2023 A framework based on OPC UA to implement a cooperative,
connected, automated mobility decision support platform.
Simplifications about standardization, security, scalability, and
governance for its practical and large-scale implementation.
[106] 2024 A video-based trajectory dataset generated from drone record-
ings intended to facilitate safety research by providing traffic
trajectories.
CitySim is an advance in generating data for road safety and
DTs, but the article does not delve into the practical challenges
for the real implementation of these DT.
VOLUME , 11
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
Efficiency
Transport SDGs
Twinned
Entity
Traffic
Environment
Transport
Network
Routes
Vehicles
Safety Ecology
[117] [114] [115]
[111] [113]
[118]
[70] [109]
[112] [119]
[107]
[110]
[108] [116]
reference {twinned entity transport SDG}
For references address more than a Transport SDG:
[120]
FIGURE 6: Main research on DT for agents’ behavior in
ITS classified by the transport SDG they address.
eration from a macro perspective. The information obtained
was used for the spatiotemporal prediction of congestion
induced by accidents, acting to mitigate the adverse effects
while responding to traffic accidents promptly. Based on the
simulation and identification of such conflicts, traffic scenes
can be simulated to support the decision-making of traffic
management centers.
Taking advantage of the DT in the almost real-time
interaction between physical and cyber entities, Fu et al.
[101] proposed a framework for the DT of an intelligent
highway. The solution’s efficiency was evaluated through a
case study based on simulations with SUMO, mapping the
driver-vehicle-roads set in a cybernetic system. The simu-
lator allows the creation of microscopic models to model
individual driving behaviors and the complex interactions
between adjacent vehicles and mesoscopic models that focus
on the heterogeneity of driver and vehicle behaviors in
probabilistic terms. Also, to improve the monitoring of road
infrastructures, Marai et al. [97] proposed deploying a system
on roads that creates a DT of the road by constantly sending
data to the Edge.
Recently, Thonhofer et al. [105] proposed a Cooperative,
Connected, Automated Mobility (CCAM) Decision Sup-
port Platform (DSP) framework, which is conceived as a
digital twin with a specific scope, clear requirements, a
versatile architecture, and mature technologies that enable
practical implementations. The framework explores a well-
defined set of requirements for information that facilitates
the development of interoperable information models for the
transportation infrastructure, focusing on the content of the
information. Using the OPC Unified Architecture (OPC UA)
platform, the authors validate the general solution concept
with different implementations and several demonstration
use cases.
C. AGENTS’ BEHAVIOR IN ITS
DT-enabled solutions have been recently proposed to meet
the challenges of urban mobility that directly influence the
operation and management of ITS, improving route selection
or conflict resolution at intersections as shown in Fig. 3c.
Chen et al. [107] developed a mobility network based on
DT to extract accurate profiles of bus stations. The data ob-
tained was used in analytical operations to perform transport
management autonomously. A co-simulation framework that
combines vehicle and traffic simulation was proposed by Shi
et al. [108]. The framework employs a deep learning algo-
rithm with video data to complete the input–model–output
full-chain autonomous driving co-simulation testing method
system.
Fan et al. [70] proposed a mobility DT to predict users’
movements based on the current urban state. This solution
predicted trajectories that respond to different conditions by
filtering or augmenting the historical database concerning
specific simulation tasks. The system encoded the daily
variation of human mobility at the metropolitan level, au-
tomatically extracting the mobility trends of the city and
then predicting long-term and long-distance movements at
an approximate level. The coarse forecasts are then resolved
at a satisfactory granularity level through a probabilistic path
recovery method, which offloads most heavy calculations to
the offline phase. Furthermore, Wang et al. [115] developed
a mobility DT defined as an edge cloud-based framework
powered by artificial intelligence to serve mobility services.
The solution consists of large blocks in physical space (i.e.,
human, vehicle, and traffic) and their associated DTs in digi-
tal space. The proposal was evaluated through a personalized
adaptive cruise control system case study. Through a cloud-
based ADAS, the vehicle DT provides visual orientation and
control commands towards the vehicles connected in a HITL
simulation.
Hasan et al. proposed PlatoonSAFE [120], an open and
detailed simulation platform to evaluate vehicle platoons’
safety and fault tolerance. Through a detailed communication
model and the integration of prediction capabilities through
ML, PlatoonSaFE allows the assignment of platoon perfor-
mance levels according to the quality of the communica-
tion. Considering the interaction between CAV and human-
driven vehicles, the platform simulates maneuvers such as
joining, lane changing, platoon formation, and dissolution
with different communication delay levels. The proposal was
used to simulate diverse scenarios with varying configuration
parameters, such as speed, inter-vehicle distance, and com-
munication quality, to evaluate emergency braking strategies
and platoon safety strategies under realistic road conditions.
12 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
TABLE 3: Summary of main research on DT for agents’ behavior in ITS.
Ref. Year Main contribution Weakness and limitations
[109] 2019 Introduce a modeling scheme for mobility based on a time series
behaviors of EVs to evaluate the charging algorithm and pile
arrangement policy.
Lacks discussion on real-world challenges, including scalability
issues, connectivity limitations, and communication constraints.
It overlooks critical security concerns, such as vulnerabilities in
physical systems and data.
[107] 2021 A DT mobility profiling framework to learn node profiles on a
mobility network, capturing the complex spatiotemporal features
in traffic scenario.
The paper does not address connectivity issues, optimized com-
munication standards, and latency concerns in cross-platform and
deterministic network integrations.
[110] 2021 A connected-corridor application for dynamic, real-time, data-
driven traffic simulation models to dynamically attain high-
fidelity vehicle record information.
It does not directly address legal considerations related to data
governance and privacy ensuring user transparency, which are
critical for legal and ethical compliance in such systems.
[111] 2021 A vehicle path planning scheme based on virtual vehicle density,
where the cloud gives different rewards to different road sections
to reduce latency.
There is a lack of discussion regarding scalability, infrastructure,
and connectivity challenges, as well as concerns about reliability,
security, and privacy.
[112] 2021 A simulation model for bus routes, to modeling processes of
several passenger services.
It fails to address the high computational costs, real-time data
integration, and the need for robust communication standards.
[113] 2021 A model to construct the bus operation chain, giving feedback
of the whole process of demand-responsive transit service.
The solution presents a specific scenario and challenges in estab-
lishing performance evaluation metrics across various domains.
[114] 2021 A intelligent vehicle behavior analysis framework that uses deep
learning and Kalman filtering to track vehicles.
Lacks on scalability and overlooks high infrastructure costs and
network challenges, key for real-world application.
[108] 2022 A data-driven co-simulation framework for vehicle dynamics,
sensors, and traffic environment modeling, using SUMO and
CARLA.
Overlooks key challenges like standardization and scalability,
neglecting optimized communication standards.
[70] 2022 A two-stage human mobility predictor that extracts citywide
mobility trends as crowd contexts and predicts long-term and
long-distance movements.
It doesn’t address challenges including connectivity reliability,
detailed security and privacy measures, scalability, and required
infrastructure.
[115] 2022 A mobility DT defined as an AI-based data-driven
cloud–edge–device framework for mobility services.
Lack of clear direction or promotion of standardization hinders
scalability and governance while amplifying dependence on
domain-specific knowledge within the proposed framework.
[116] 2022 A DT-assisted real-time traffic data prediction method by analyz-
ing the traffic flow and velocity data monitored by IoV sensors
and transmitted via 5G.
Faces key challenges related to scalability, infrastructure, and
connectivity, particularly regarding the assumptions about the 5G
environment.
[117] 2022 Technical methods of DT for urban transportation process in-
tegration: creating architecture, analyzing system function, and
digital technology integration.
It neglects aspects such as scalability, security, and governance.
[118] 2022 An anticipatory algorithm for a real-world dial-a-ride problem
that uses a DT framework to analyze and optimize the algorithm.
Fails addressing challenges related to security, privacy, and
governance in these complex systems. It misses crucial practical
issues needed for successful deployment.
[119] 2023 Uses a genetic algorithm to understand the bus dynamics. It neglects aspects such as scalability, security, and governance.
[120] 2023 PlatoonSAFE, an open-source simulation tool designed to eval-
uate the safety of vehicle platoons.
Lack of attention on human factors, commercial implications,
and governance required for the successful adoption.
Obtaining real-time information on the actual traffic flow
is another fundamental aspect of ITS. Hu et al. [116] mod-
eled the traffic flow for creating a DT for real-time traffic pre-
diction according to speed and traffic flow data monitored by
distributed traffic cameras. Collecting data in large cities to
understand the impact of transportation services on transport
has been one of the main tasks of transport researchers and
engineers. By analyzing traffic data supported by DTs, traffic
managers can optimize traffic scheduling and formulate
transportation policies. The DT of the transportation system
of a medium-sized city was proposed by Khalil et al. [121],
considering several modes of transportation, such as public
transportation, shared transportation services, and private
vehicles. This DT evidenced new opportunities, creating an
environment for policy learning with reinforcement learning
using open data sources and agent-based simulations. A
DT was also proposed by Abhilasha et al. [110], which
leverages real-time data streams to model the current traffic
state and provide dynamic feedback on traffic. In this case,
the robustness and feasibility of the proposed DT architecture
were demonstrated using a testbed. The testbed allowed
the injection of real-time control signals and volume data
VOLUME , 13
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
flows in the traffic simulation model to evaluate the dynamic
performance metrics.
Choosing the best urban traffic route, shortening travel
time, and alleviating traffic pressure are other critical aspects
of transportation systems. Thus, a route planning scheme
DT-enabled was proposed by Hui et al. [111] to facili-
tate traffic management, considering the personalized users’
demands. In the solution, DT vehicles interact with each
other from the cloud to make route planning decisions in
advance. According to the traffic density of the different
road sections, rewards were established to encourage the
vehicles to obey the scheduling instructions. Moreover, Wang
et al. [117] presented an urban transportation project where
the technical methods of the DT of the urban transport
process are addressed, making the architecture, analysis of
the system function, and integration of digital technology
available for such projects. As a case study, the best path
system design method was provided by building the frame
structure. The field detection module was used to perform
the virtual driving test and the detection data processing to
ensure route selection accuracy.
For the electric vehicle (EV) industry, the driving experi-
ence is highly dependent on the availability and accessibility
of vehicle charging infrastructure. As the number of charging
piles increases, carefully designed arrangements of resources
and efficient infrastructure utilization are essential for this
industry’s future development. Zhang et al. [109] proposed
a DT by modeling the mobility based on the behaviors of
a time series of EVs to evaluate the charging algorithm
and the grid layout policy. The DT EV behavior and route
choice are dynamically simulated based on time-varying
driving operations, travel intent, and charging plan in a
full-scale simulated charging scenario. With a DT mobility
model, the EV behaviors and interactions were simulated
to study the efficiency and quality of charging on both
the supply and demand sides. Furthermore, the performance
of different proposed charging scheduling algorithms and
charging stack deployments was evaluated by simulating
traffic and charging on a large scale of inter-connected EVs.
Several DTs-powered solutions have been proposed for the
public transport sector to improve route planning. Creating
DTs of real bus routes requires the development of suitable
approaches for modeling individual processes. Zhukov and
Moroz [112] proposed a simulation model of bus routes
aiming to clarify the approaches to model the processes of
getting on and off passengers, the parking time of vehicles
at a stopping point for passenger exchange, and the waiting
time of passengers to board the vehicle. To address demand-
sensitive traffic (DRT) problems, Deng et al. [113], with
the support of DT, performed the representation of a DTR
system to explore essential details of this system. The pro-
posal describes the bus transit process between two stations,
suggesting that DT-based solutions contribute effectively to
the construction of the travel chain for the traveler.
In trajectory calculus problems, the performance of the
different anticipatory algorithms when the dynamism in the
system increases is also a fundamental question. Ritzinger
et al. [118] proposed a framework based on DTs, al-
lowing sophisticated algorithm performance analysis and
re-optimization strategies. Special attention was given to
the communication and data synchronization between the
real environment and decision-makers, which is vital when
dealing with dynamic vehicle routing problems. The DT-
based module implements a state machine where all possible
vehicle states and transitions are defined, and computa-
tional experiments are performed on a set of test instances
generated based on information and data from the real
environment. Similarly, Li et al. [114] proposed a behavior
analysis framework for intelligent vehicles based on a DT
that supports the detection of abnormal behavior. In the
solution, the tracked vehicle was assigned to a DT virtual
scene, and the behavior of each vehicle was tested according
to custom detection conditions configured in the scene. The
solution in [114] was an innovative and efficient scene-
building toolchain and production process, integrating high-
definition mapping, data collection, photogrammetry, and
process generation technology.
D. INTERNET OF VEHICLES
The Internet of Vehicles (IoV) leverages advancements in
sensing and communication to build an environment in
which vehicles are conceived as (i) powerful multi-sensor
platforms; (ii) embedded computational features; and (iii)
elements capable of communicating with other vehicles and
road infrastructure as shown in Fig. 3d. With the potential
to deploy compute-intensive applications, edge computing
is combined with IoV powered by DTs to enhance in-
telligent transportation capabilities. Furthermore, the need
for in-vehicle computing resources can be supplemented
by upgrading vehicle DTs and offloading services to edge
computing devices.
Xu et al. [122] implemented a multi-user download system
where the QoS is reflected through the services’ response
time. The approach applied a service offload method with
deep reinforcement learning for IoV powered by a DT. With
this method, performing comparative experiments with real-
world service data is feasible to assess the efficiency and
adaptability of service offloading. Xu et al. [123] also applied
a DT in an ITS to simulate the offloading strategies, building
a virtual representation of the system that examines the
states of the roadside units. Zhang et al. [124] proposed a
new vehicular edge computing network to create an edge
management framework that improves multi-agent learning
efficiency through DT technology while improving replica-
bility performance between virtual and physical networks
through a learning approach. A distributed multi-agent learn-
ing scheme minimizes the download cost of vehicular tasks
under strict delay constraints and dynamically adjusts the
state mapping mode of the DT network.
14 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
To meet ultra-low network latency demands of in-vehicle
internet applications, IoV enables edge devices to share their
communication, computing, and storage resources through
intelligent edge cooperation. However, efficiently allocating
such resources using machine learning techniques and arti-
ficial intelligence requires a large amount of training data
and high computing power that is impossible to achieve
in a resource-constrained onboard unit or a roadside unit.
Liu et al. [125] proposed a DT-enabled intelligent edge
cooperation scheme, enabling optimal resource allocation
and intelligent edge cooperation. The proposal aims to min-
imize the delay to meet the requirements of time-sensitive
applications in ITS. A detailed analysis of the computational
and communication resources between the physical world
and the virtual space was carried out by introducing the
DT technology. Through this DT, network resources are
managed and allocated from a global perspective, which
promotes collaboration between edge servers and improves
the utilization of idle system resources effectively.
Yuan et al. [126] also proposed a DT-powered framework
designed for IoV, focusing on vehicular task offload func-
tions. The DT of the network devices is mapped in real time
to represent latency states and power consumption. Since
vehicular networks’ latency and power consumption come
from wireless communication and computing offloading, a
wireless communication model was developed to optimize
offloading decisions and thus minimize latency and power
consumption. The solution was implemented in an IoV
assisted by MEC to improve the quality of communications
and guarantee quality of service (QoS). Lastly, Liu et al.
[127] proposed a DT transport system based on virtual reality
for the IoV. In the solution, the basis of the vehicle in the
simulation system is the traffic flow information collected
in the real traffic scene. The objective was focused on the
collaborative control of vehicles at traffic intersections based
on V2X communication to perform collaborative decision-
making of multiple vehicles and guarantee vehicle safety.
In heterogeneous vehicular networks (HetVNets), base
stations can exploit the massive amounts of data collected
by vehicles to complete federated learning tasks. Hui et
al. [128] proposed a scheme enabled by DTs for multitask-
ing federated learning in HetVNets. For this, the training
capabilities of the base station are analyzed, considering the
available training data, the declared price, and the training
experience. The task requesters and the base stations create
their DTs in the cloud, where a game-based scheme attempts
to achieve efficient matching between the requesters and
the base stations in the DT networks. Also, in HetNets,
optimizing the connection between the vehicle and its DT
throughout the vehicle journey is challenging due to the
uneven distribution of vehicles in the networks and the
dynamics of heterogeneous wireless links. To address this
problem, a learning-based heterogeneous network selection
scheme in DT-empowered IoV was proposed by Zheng et
al. [129]. The proposal jointly considers the dynamics of
wireless conditions, vehicles’ mobility, and heterogeneous
access links’ characteristics. As a result, vehicles can reuse
the knowledge they have learned from previous tasks to
solve new tasks faster or use better solutions. New vehicles
incorporated into the network can also use the experience of
expert vehicles to implement their knowledge system rapidly.
E. DISCUSSION: A CONTEXT ANALYSIS
By revisiting RQ #2 (“What can DTs represent from ITS’
new physical environments?”), we conclude that DTs can
represent almost all the elements of the ITS ecosystem,
including autonomous and connected vehicles, transport
infrastructures, and behavior of ITS agents, boosting ITS
applications closely related to the fulfillment of the SDGs. In
numbers, from the total papers evaluated in Section III to re-
spond to the RQ #2, DT-based solutions for autonomous and
connected vehicles represent 43%, intelligent infrastructure
applications 27%, and applications concerning the agents’
behavior in ITS 30%. Even though DT-ITS solutions are
being used to address the main areas of ITS, it is evident
that the automotive sector leads this section with a strong
commitment to using DT in vehicle manufacturing. Besides,
regarding the adoption of DT-ITS to meet the demands of
the SDGs, 53% of the works address problems of improving
efficiency in the use of ITS resources, safety problems are
addressed in 67%, and solutions that consider the ecological
aspects of ITS 10%. The low interest in solving ecological
problems in ITS using DTs is striking. Only five publica-
tions address ecological issues in some way, and in most
of them, the environmental approach is a collateral result
of research focused on efficiency or safety. This behavior
should be understood as an alert for DT-ITS researchers and
practitioners to address ecological problems in ITS solutions
more responsibly. In numbers specific to each ITS area,
safety studies lead the CAV area with 91% of the works,
while ITS efficiency studies lead with 80% the area of ITS
agent behavior. As mentioned above, the SDGs are addressed
more broadly for intelligent transportation infrastructure,
with ecological problems being addressed less. Finally, it
should be noted that starting in 2021, publications of research
work and projects addressing DT-ITS solutions will increase
considerably.
F. DT-ITS REAL-WORLD STUDY CASES
To bring evidence that precursor technologies, discussed in
Section II, are being integrated toward creating the first
prototypes of DT-ITS, here we exploit also initiatives toward
real-world implementations based on sensing consolidation
and scenario simulation, as well as test systems using in-
the-loop approaches. The twinned entities, as discussed in
Section III, are indicated for each study case in Table 5
alongside the tackled SDGs. However, many DT-based appli-
cations aimed at transport systems are usually part of more
complex urban DT approaches. These solutions enhance city
VOLUME , 15
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
TABLE 4: Summary of main research on DT for IoV in ITS.
Ref. Year Main contribution Weakness and limitations
[122] 2022 A service offloading method using deep reinforcement learning
to address the issue of edge computing devices becoming over-
loaded due to excessive service requests, which deteriorates the
QoS.
Practical limitations and simplified assumptions of the model.
It is assumed that the vehicle’s DT is constantly updated with
the physical vehicle without considering data inconsistencies and
latency in the updating of DT.
[123] 2022 A comprehensive approach that combines DTs, decision theory,
computation offloading, and service caching to improve the per-
formance of ITS by minimizing latency and optimizing resource
utilization.
Lack of discussion about the practical limitations related to in-
frastructure and security. The validity of the model’s assumptions
requires greater attention to ensure feasibility and applicability.
[124] 2022 A framework that improves the learning efficiency of multiple
agents through DTs for efficient resource management in vehic-
ular edge computing networks, minimizing task offloading costs.
The practicality and assumptions about infrastructure and con-
nectivity, as well as limitations in security and privacy.
[125] 2022 The development of a DT resource allocation scheme to improve
edge collaboration in the IoV environment minimizing latency
in task processing through the efficient allocation of communi-
cation, computing, and storage (3C) resources.
Limited discussion related to high-capacity infrastructure, reli-
able connectivity, scalable security solutions, and the validity of
model assumptions to ensure real-world viability and applicabil-
ity.
[126] 2022 An approach to optimize resource allocation for task offloading,
taking into account a model of the IoV environment, helping in
reducing latency and energy consumption.
The lack of specific security, scalability, and connectivity so-
lutions limits the system’s practical application. lack of well-
defined business models and governance frameworks.
[127] 2022 Integration of VR and the IoV to develop a real-time traffic
management system to generate accurate traffic models, efficient
task allocation, collaborative vehicle control, and data privacy
protection.
Lack of a comprehensive evaluation of scalability, costs, inter-
operability, and human factors. Lack of consideration for real-
world complexities and its dependence on idealized assumptions
that limit practical applicability.
[128] 2022 A framework that uses DTs to optimize the allocation of
federated learning tasks in heterogeneous vehicular networks,
considering both the task requesters’ personalized needs and the
roadside units’ differentiated capabilities.
Lack of a comprehensive assessment of scalability, costs, interop-
erability, and human factors. The complexities of interoperability
between different systems and devices are overlooked. Simplifies
traffic dynamics by assuming predictable user behavior.
[129] 2022 A learning-based heterogeneous network selection scheme for
data synchronization between vehicles and their DTs to improve
learning efficiency and reduce data synchronization latency.
Lack of consideration for real-world complexities, specifically
regarding scalability, infrastructure, connectivity, security, and
governance.
transportation systems and address other domains of smart
cities, such as civil infrastructure, energy, and security.
Among the early initiatives showcasing real use cases of
DT-ITS, it is worth mentioning the prototype of Herrenberg
in Germany [130]. There are two modules focused on the
transportation system, which utilize a street network model
based on the theory and method of space syntax, along
with an urban mobility simulation. The street network model
enabled the analysis of movement potentials for both cars
and pedestrians, helping to determine how ”central” a street
segment is relative to others within the urban system. Fur-
thermore, the space syntax method enhanced understanding
of the street network configuration, which pertains to the
underlying logic of publicly accessible places in the urban
system. The model was further developed using the SUMO
traffic simulation to gain deeper insights into traffic behavior.
In another similar initiative, the Department of Transport in
Seychelles implemented a DT of the road transport system
on the main island, Mah´
e [131]. This project evolved using
existing reports and drone monitoring to build and calibrate
a reliable SUMO microscopic traffic model that reflects
the current traffic situation. The DT was also employed in
the city of Victoria to assess the impacts of the planned
pedestrianization on the road transportation system.
Several use cases focused on transport systems have
emerged from major European projects, but still in the
context of urban DTs. One of the most significant initiatives
is the Digital Urban European Twins (DUET) [12], which
has led to implementing urban DTs in various regions across
Europe. Among the key applications, the city of Antwerp in
Belgium has implemented CityFlow, a DT that utilizes vari-
ous real-time data sources to better capture the multimodality
of city traffic. The CityFlow model can estimate traffic
density by type, providing valuable information for stake-
holders involved in decision-making and the development of
public policies. Another DT of the transport network has
also been implemented in Flanders, Belgium, and Athens,
Greece, to evaluate the impact of certain changes in urban
circulation plans on low-carbon zones. Transport network
analysis, which calculates data on traffic volume and the road
network, can also be utilized to build air quality emission
models. In this context, the city of Pilsen employs a DT to
monitor local road transport features such as traffic volumes,
16 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
TABLE 5: A list of DT implementations in transportation systems showcasing case studies.
Case Study Transport SDGs Twinned Entity Main Contribution
Herrenberg, Germany [130] Efficiency,
Ecology
Transport network A street network model to analyze movement potentials for both cars
and pedestrians for an optimized design of public spaces in the city.
Victoria, Seychelles [131] Efficiency Transport network A reliable microscopic traffic model that reflects the current traffic
situation to assess the impacts of the planned pedestrianization.
Antwerp, Belgium [12] Efficiency,
Ecology
Traffic environment Citizen-centric DT that use different real-time data sources to better
capture the multimodality of city traffic.
Flanders, Belgium and
Athens, Greece and Pilsen,
Czech Republic [12]
Ecology Transport network Evaluate the impact of certain changes in urban circulation plans to
build air quality emission models and the impact of local road transport
features on noise pollution.
Chattanooga, USA [11] Efficiency,
Ecology, Safety
Traffic environment A comprehensive DT platform that integrates data, simulations, and
cyber-physical control to optimize urban mobility, as well as a modular
design that allows its adaptability and generalization.
London, UK [132] Efficiency Transport network A tactical highway traffic assignment model used to assess the impact
of various schemes and evaluate mitigation strategies.
Uzbekistan [13] Ecology, Safety Highway Develop a comprehensive and innovative DT-based approach to road
pavement design.
vehicle types, and traffic conditions to assess their impact on
noise pollution.
One of the most comprehensive use cases for trans-
portation system applications is the Chattanooga Digital
Twin (CTwin) [11]. This solution optimizes the city’s ur-
ban transportation systems by reducing traffic congestion,
incidents, and vehicle fuel consumption. CTwin connects
a network of diverse IoT sensors, transportation infrastruc-
ture, and third-party data services to gather and integrate
large volumes and varieties of urban mobility-related data.
Using this data in a cloud computing environment, CTwin
executes advanced simulation models to produce analyses
and optimized strategies for mitigating traffic congestion.
Through simulations, CTwin can construct continuous traffic
flow using discrete roadside sensors, implement a real-time
signal control algorithm, and evaluate the performance of
transportation systems.
For transportation infrastructure, many agencies world-
wide are enhancing operational efficiency and optimizing
overall performance with the support of DTs. Transport
for London (TfL) has been proposed for exploring the
concept of a DT to enhance urban mobility and improve the
efficiency of its transport network [132]. To this end, TfL
has developed the Operational Network Evaluator (ONE), a
tactical highway traffic assignment model used to assess the
impact of various schemes and evaluate mitigation strategies.
The ONE serves as a virtual representation of real-world
road traffic conditions in the London area. TfL extensively
utilizes DTs to assess the impact of new initiatives, including
cycle routes and major road redesigns. It has also aided in
the operational analysis of road and river crossing closures
and their effects on bus journey times. Furthermore, TfL
has implemented a DT solution for the underground system,
bringing visual intelligence to line tracks, stations, and
depots [133]. This solution mapped the geometry of under-
ground spaces and detected environmental pollutants. These
readings helped establish a foundation for measuring carbon
emissions and were crucial in tracking progress toward
environmental goals. A DT provides real-time insights and
enables data-driven decision-making to optimize operations,
ensure worker safety, and minimize environmental impact.
A project in Uzbekistan aimed at upgrading part of a high-
way utilized AI and DT to support greener and more resilient
transport infrastructures [13]. This use case demonstrates
how road design can be significantly improved by making
informed, data-driven decisions at an early stage based on
local conditions. Specifically, the DT platforms were used
to develop a comprehensive and innovative approach to
road pavement design. This approach supports quick due
diligence and facilitates investments in lower-carbon, safer,
circular, and more resilient road networks. The DT was
integrated into the ORIS platform 2, incorporating data from
real environmental characteristics, including local sourcing,
traffic, and weather conditions. By using the DT, the project
team was able to assess multiple pavement design options,
analyzing the initial pavement design in terms of costs,
carbon emissions, resilience, and resource consumption.
Moreover, the simulation results allowed for the evaluation
of designs and alignments (both vertical and horizontal),
providing recommendations on road safety by considering
geometry, alignment, visibility, and traffic.
IV. DT-ITS REFERENCE FRAMEWORK
An important observation about the surveyed literature pre-
sented in the previous sections is the lack of architectural
frameworks to develop DT for sustainable applications in
modern ITS. Therefore, to summarize our findings and
address RQ #3, this Section collects elements to propose a
2https://www.oris-connect.com/
VOLUME , 17
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
Physical Models Data-driven Models
DT-based
ITS Models
Model Layer
Application Layer
Data Layer
ITS Data Lake
Simulation Learning Prediction Visualization
APIs APIs
Communication
Technologies
Computing
Platforms
Static Data Dynamic Data
IoT/Telematics
Physical Layer
Traffic
Agents Intelligent
Infrastructure
Traffic Behavior
integration
combination
Analytic Apps
CAV
Intelligent
Infrastructure
Other ITS
Applications
Behavior of
ITS Agents
ITS Apps
Sustainable
Transport Goals
Safety
Ecology
Efficiency
Safety
Efficiency
FIGURE 7: DT-ITS reference framework.
concise reference model for DT-ITS. In addition, discussions
on DT-ITS functional and non-functional requirements are
also presented. Figure 7 illustrates a systematic framework
to accurately and timely develop DT-driven sustainable appli-
cations in transportation systems. The framework consists of
4 layers and considers the ITU-T Y.dt-ITS recommendation
draft [48]. Each of the layers and the interaction mechanisms
between them are presented below.
1) The physical layer
This layer represents the elements of the transport ecosys-
tem and the processes of combination and integration. The
physical layer includes all the transport items to realize a
full-cycle operation and management process of intelligent
transportation. Thus, this layer is in charge of the cutting-
edge processes in the operation of any DT-driven application,
namely, sensing and actuation. Dynamic behaviors, events,
and processes are detected and synchronized to higher layers
to create DT-based models. The physical space is represented
through various functional blocks.
The intelligent infrastructure represents the first functional
block, which includes design tasks, physical infrastructure
maintenance, and system operation management with a
macro approach. The monitoring processes use specific
sensors for the civil infrastructure, and others are specific
to the transportation area for operational management. The
actuation processes are an essential component of human
intervention, especially in cases directly related to infras-
tructure design and maintenance.
The second functional block comprises the traffic agents.
This block includes vehicles (e.g., connected and au-
tonomous vehicles, assisted driving, and conventional). For
them, the most used sensors are the geolocation modules,
perception sensors (e.g., LIDAR, radars, cameras), and the
information from the intra-vehicle network. The action fo-
cuses on the greatness of vehicle control (e.g., steering
system, acceleration, brakes, and others) in the case of fully
autonomous vehicles or with some level of autonomy. In
contrast, in the case of conventional vehicles, the action
is carried out by the drivers. Non-vehicle agents, including
drivers, pedestrians, and cyclists, also known as vulnerable
road users (VRU), are also represented in this block. The
sampling processes for these participants are carried out
actively through the human-machine interface and passively
with in-cabin sensing, environment sensing, or wearable
devices. The action processes for non-vehicle participants
are fundamentally aimed at drivers through driving assistance
systems, which can act more precisely and thus influence the
entire transport system. Other action methods for pedestrians,
cyclists, or passengers are used on the roads through their
wares or signaling systems.
The last functional block deals with traffic behavior and
other behaviors that act directly on the transport system, such
as the mobility characteristics, the drivers’ profile, and the
actions of local regulatory agencies. The sensing in this block
is closely related to information obtained by the previous
functional blocks in those cases where the information of the
transport environment is used to compute traffic and mobility
information. Others depend exclusively on the relationship
between traffic management companies, regulatory agencies,
and authorities to keep up-to-date with policies and practices
specific to each scenario.
Several communication technologies are used to synchro-
nize the physical transportation world with their virtual
entities. Similarly, DT-based models generate signals or
instructions through actuators within their platform to trigger
physical behavior [98]. Transportation-based scenarios uti-
lize wireless technologies for communication between partic-
ipants, agents, and the infrastructure. Considering 2030 as a
landmark for sustainable mobility, B5G/6G networks need to
pave the way for safe, affordable, accessible, and sustainable
transport systems and improve road safety. Recently, the
large volumes of data sensed in transportation scenarios,
primarily driven by the use of computer vision applications
for detection, classification, and tracking of transportation
18 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
system agents, have established technical requirements for
which two main solutions of radio access technology have
been adopted: B5G communications and Wi-Fi [134].
Due to the importance that transportation system commu-
nications have acquired in recent years, particularly in the
context of autonomous vehicle communications, 3GPP rec-
ognizes V2X communications as a primary service case for
5G/6G networks. 5G NR C-V2X builds upon the powerful
features standardized for 4G and 5G. Thus, 5G NR V2X
is being designed to support V2X applications with varying
degrees of latency, reliability, and throughput requirements
[135]. The standard considers the use of mmWave bands for
V2X applications, particularly those requiring short range
and high to very high throughputs. The 5G NR V2X can
provide unified support for all V2X applications and must
be capable of supporting not only advanced V2X applica-
tions but also basic safety applications that present-day C-
V2X supports. Recently, a technical report from the 5GAA
Automotive Association [136] specified that among the main
expectations and priorities of road operators is for telecom
providers to offer robust and reliable C-V2X services to
support road traffic modeling and planning, as well as the
provision of safety information to road users via DT. It is
also recognized by the technical community that these C-
ITS technologies are fundamental for enhanced services for
real-time traffic monitoring and control included in DT-ITS.
On the other hand, Wi-Fi has been part of the planning
in the automotive industry for some time [134]. Wireless
Access in Vehicular Environments (WAVE) is defined in
the IEEE 802.11p amendment to support Intelligent Trans-
portation System (ITS) applications. The 802.11p standard
can facilitate communications for mobile elements traveling
at relatively high speeds, such as vehicle-to-vehicle (V2V),
vehicle-to-infrastructure (V2I), and V2X communications.
However, the adoption of Wi-Fi for transportation system
communication has been limited due to the poor scalability
of the 802.11p standard in high-mobility environments. To
address these challenges, work has begun in recent years on
the new 802.11bd standard as an evolution of 802.11p for the
next generation of V2X communications [135]. Along with
5G NR V2X, 802.11bd is expected to support more advanced
V2X applications with stricter QoS requirements. The stan-
dard aims to facilitate communications with mobile nodes at
speeds of up to 500 km/h, achieves twice the communication
range of 802.11p, and provides vehicle positioning with a
location accuracy of up to 1 meter.
2) The Data Layer
The data layer is based on collecting real-time data from
the sensory and sampling processes produced in the physical
layer, which, combined with historical data from the system,
provide a sufficiently complete data lake for constructing the
models used by the DTs [87]. For this, it is recommended
to collect the most accurate data possible and more abun-
dant information, giving the DT features reflecting the real
environment state. Depending on the processes or physical
entities that are being monitored, the data may have i) a
dynamic perception of the physical space concerning the
operations of the elements of traffic participants and their
behaviors; or ii) a static nature more linked to the description
of the environment and the infrastructure [137]. Dividing the
system’s static information and the various situations from
dynamic elements of the system provides a logical basis
for reasoning about the system’s behavior, which can be
represented more accurately in the data model used by the
DT in its prediction, analysis, and visualization tasks.
In the context of ITS, the proliferation of mobility data and
diverse data formats from roadside equipment—transferred
using various protocols and processed either at the edge or
in the cloud with advanced AI capabilities—along with the
emergence of connected and automated vehicles, introduces
challenges for the DT-ITS framework. Therefore, this layer
is responsible for detecting and classifying structural, non-
structural, and semi-structural data to integrate and distribute
data of different types from various sources, enabling inter-
vention and data entry [138]. To this end, the description
of the data elements in the DT-ITS construction process
is carried out in this layer, covering both the syntax and
semantics of the data. This includes a definition of the format
and content of the data, as well as a description of how to
use these data elements. All data structures supported by
the DT-ITS must be specified in a communication-agnostic
manner to enhance the interoperability of the system.
3) The Model Layer
The data lake is used to create computational models that
describe the infrastructure, the agents participating in the
traffic, and the operational processes of the ITS. A data
processing stage is necessary for constructing these models,
which is carried out in this layer to create physical and data-
driven models [58]. Combined, these two approaches permit
the provision of a prognosis service that predicts the future
parameters of a transportation system. As such, physics-
based models are used to preprocess data coming from
multiple time-scale data streams, while the ML models at the
upper layer extract information to predict future behaviors
of each time series parameter. Likewise, a stage of post-
processing the data and combining different sets of data and
models allows for obtaining the most complex DT models
(decision models) that provide a complete understanding of
the data and the physical entities and processes. Once the
DT model is built, it receives real-time information from the
transport system to stay updated and more accurate. There
is feedback to the physical layer as the DT is updated from
the simulation and analysis, modifying the physical entities
and processes. In this way, the DTs go beyond the simple
conceptual design of ITS applications because the DT-based
models are built to act as a simulator of reality. Still, they can
remotely modify and manage any physical entity or process
in transportation systems. The DT-based model interacts with
VOLUME , 19
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
the application layer through APIs implemented for different
external applications to consume the model information.
When discussing information modeling, the common prac-
tice is to create an ontology model written in a language that
supports semantic richness, enabling a greater capacity for
exchanging machine-readable content [139]. The fundamen-
tal concepts of ITS operations that address the SDGs can
be defined clearly and unambiguously through the formal
description of entities and their relationships. Thus, a well-
defined domain ontology at this layer enables the accurate
representation of physical entities for the DT-ITS. Ontologies
are necessary to provide semantics to the data collected at the
Data Layer. However, creating data models from scratch can
be an inefficient approach for prototyping, considering that
there are already solutions available to support the creation
and operation of DTs, which are typically referred to as DT
platforms.
The Platform-as-a-Service (PaaS) for developing DTs,
especially for DT-ITS, allows for efficient data collection
mechanisms in road intersection scenarios, storing the in-
formation in unified and efficient data repositories. Most of
these platforms, such as Azure Digital Twins, Eclipse Ditto,
or FIWARE [140], enable the design of DT architectures
that integrate IoT devices and connect to brokers for real-
time data exchange. A key feature of these platforms is the
ability to define custom vocabularies using domain-specific
languages (DSLs) to provide comprehensive DT-ITS solu-
tions. A significant advantage is the implementation of these
containerized format solutions, which makes them highly
scalable and easy to implement and replicate. Furthermore,
microservices facilitate communication and data manage-
ment, allowing integration with other backend systems via
well-known APIs.
4) The Application Layer
This layer is responsible for computing two types of applica-
tions. Analytics applications of the DT that use the models
for simulation and analysis, learning tasks of intersection
patterns, urban mobility, and visualizing current conditions
of the transport system [115]. For most of these cases, the DT
and their applications must work offline to explore different
scenarios of the ITS without being tied to the elements of the
physical world. The analysis tasks allow an understanding of
the crucial issues of the transport system and its processes,
analyzing the real world through the DT-based model. In
complex systems such as ITS, when system assets interact
with each other, many unexpected situations can arise that
are almost impossible to analyze by operation centers or by
simple system simulations. Through the DT-based model, the
applications help to optimize interactions to obtain unknown
information and assist in decision-making.
This ability to understand the physical assets of the
transportation system, as well as its dynamics, is achieved
through the use of artificial intelligence (AI) and machine
learning (ML) techniques, which could be considered the
brain of the DT-ITS [141]. Since both AI and DT systems
require data to function, it is a logical step to integrate them.
At this point, the choice of ML models to be applied for each
transportation application depends on the use cases and the
modeling type of the specific DT-ITS. Since ML algorithms
are closely related to the resolution of optimization problems
and predictive analysis, a common approach is to employ
data-driven models at the Model Layer to minimize or
maximize a given process parameter. As found in most of the
works surveyed here, DT-ITS derives significant advantages
from applying ML techniques. Utilizing vast historical data
and relevant AI algorithms allows it to enhance the accuracy
of its predictions. This is achieved at this layer by leveraging
the analytical power of ML, which in turn utilizes the
information and patterns embedded in the transportation
system data to refine simulations, analyses, and ”what-now,”
”what-next,” and ”what-if” predictions.
Here, deep learning stands out from other methods for
its application in DTs, as the feature selection process is
automated by a general-purpose learning procedure without
any human involvement [142]. Deep learning has a great
capacity for modeling spatial and temporal dependencies,
making it a key tool for addressing the dynamism of DT-ITS
in detection, diagnostics, and prognostics processes. In this
way, the use of deep learning also circumvents the practical
constraints of implementing DT-ITS, where many agents
are involved, making the implementation significantly more
complex since each physical asset requires its own physical
model. Thus, using deep digital twin (DDT) concepts [143]
allows for obtaining a digital representation of the expected
behavior of a real-world transportation system and its agents
directly from operational data using a deep generative model.
Once such a representation is obtained, it becomes possible
for the DT-ITS to simulate ITS responses by sampling
from the DDT. All these capabilities using a deep learning
approach contribute to efficient transportation management
and could support other applications, such as navigation
services, route planning, and traffic control.
Furthermore, ITS applications of particular solutions for
each transportation scenario are also included in this layer.
These applications aim to address the main SDGs related
to transportation systems. Such applications consume the
model’s data together with the results of simulations, analy-
ses, and predictions for developing specific ITS applications
for monitoring and managing the various facilities of the
transit system and the transportation infrastructure. The DT
solutions can share analytical results across visualization
platforms through these applications. With them, the author-
ities responsible for the transportation system can detect and
isolate failures in physical infrastructures, diagnose and solve
problems, and recommend corrective actions. Finally, these
externals can also peer with the business layers, like in the
well-known framework Reference Architectural Model for
Industry (RAMI) 4.0 [27], to automate communications with
corporate clients, contractors, and service providers.
20 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
A. FUNCTIONAL REQUIREMENTS FOR DT-ITS
1) Efficient data collection
DT-enabled ITS are required to support bulk system data col-
lection. DT-ITS must support efficient data collection tools
and methods to collect data on all physical characteristics of
the transport system. It is desirable that data collection meth-
ods offer all data in time series or that some data aggregator
entity creates a time series from the data obtained. During
the data collection stage, it is essential to use various tools
required for the DT-ITS to support the collection of different
types of transport data with different characteristics. Some
data requires a higher frequency of collection (i.e., signal
status, agents’ location), while other data requires a higher
level of real-time (i.e., traffic flow status, agents’ location). It
is recommended that the DT-ITS explore new data collection
technologies considering the requirements of the transport
applications implemented by the DT. At the same time, such
data collection methods should be as lightweight as possible
to reduce the occupation of network and computing resources
so that it does not affect the operation of other transport
system features that use data sharing. Thus, data collection is
required to improve execution efficiency, reduce computing,
storage, and communication costs, minimize redundant data
collection, and use data compression when possible.
2) Efficient and unified data repository
With the transport system data collected, the DT-ITS must
build a unified repository to store and manage such data. The
repository must be able to extract, transfer, and upload the
data collected from the ITS physical entities and processes
and store the data. It is recommended that heterogeneous
databases be used to store the data collected from multiple
structured and unstructured data. The unified repository is
required to have the ability to store various types of system
data, including interim data and operational data. The DT-
ITS unified data repository is required to have the ability to
support real-time data acquisition and access to support time-
critical applications. In addition, the repository must provide
a unified interface for exchanging data from the transport
system with the data models created for the DT-ITS. Lastly,
the data repository is required to have the capacity to
efficiently manage massive amounts of data from complex
transport scenarios, guaranteeing the accuracy, consistency,
integrity, and security of the data.
3) Unified data models
The DT-ITS is required to define and create unified data
models for various sustainable transport applications. The
unified data models must be able to model the entire trans-
port system, the agents participating, and the interactions
between them. In addition, the data model must be designed
to fully use the unified data repository to create various
models for analysis, emulation, diagnosis, and prediction for
specific application scenarios. To do this, the unified data
model must have an interface to obtain the requirements of
the ITS applications and report the simulations’ results to the
applications. For more complex scenarios, the data model is
required to have the ability to provide services efficiently
through a combination of multiple models. Another data
model requirement is the ability to emulate and iteratively
optimize transport applications that use such a model.
4) Open and standard APIs
The DT-ITS require open and standard interfaces for infor-
mation exchange between the physical system and the DTs.
Specifically, a southbound interface between the physical
ITS and the DT data models and a northbound interface
responsible for information exchange between the DT-ITS
and transportation applications are necessary. The correct
implementation and use of such interfaces in the DT-ITS
contribute to avoiding blocking the hardware or software
provider and achieving the complete interoperability of the
DT-ITS. Both interfaces should have high extensibility to add
more features with limited parameter changes and backward
compatibility. They must also be interfaces that are easy to
access and use to handle massive data and high concurrency
and provide secure and reliable information exchange mech-
anisms.
Southbound interfaces are required to collect data from
ITS scenarios. Southbound interfaces must be able to im-
plement various collection methods, including passive and
passive collection and on-demand collection. The diversity
of data that can be collected in the ITS demands that
the southbound interfaces support several speed options to
accommodate different data requirements from applications.
In addition, these interfaces are required to deliver control
signaling that allows updating the physical system depending
on the outcome of specific applications. On the other hand,
the northbound interface must be able to deliver the require-
ments of transport applications to the DT-ITS data models.
Through the northbound interfaces, digital copies of the DT-
ITS and data models are provided to third-party applications.
In addition, these interfaces must be capable of reporting the
results of executing the data models in the DT-ITS.
B. NON-FUNCTIONAL REQUIREMENTS FOR DT-ITS
1) Compatibility
The DT-ITS need sufficient compatibility to apply to various
transport scenarios, even managed by multiple regulatory and
control entities, which may eventually be interested in vari-
ous transport applications. It is necessary to support different
transport scenarios containing physical and virtual devices
to apply DT-ITS in numerous applications. Additionally,
DT-ITS requires backward compatibility so that all updated
or new functionalities (i.e., interfaces, data repository, data
models) can work seamlessly with the functionalities of
previous versions. The DT-ITS must work with the current
transport system management implemented by transportation
operators, mainly when the DT-ITS is used for operations
and maintenance in complex transport scenarios.
VOLUME , 21
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
2) Scalability
The DT-ITS must be scalable enough to support large-
scale transport scenarios. The DT-ITS must be able to build
a virtual twin of the large-scale physical system, taking
into account, during the design and implementation stages,
the complexities that a large-scale solution introduces in
software design, data modeling, and embedding of new
features. Consequently, it must be able to scale the virtual
twin transport systems automatically according to the growth
or reduction of the physical systems. Even in cases where the
physical system is extended and, therefore, its complexity,
the DT-ITS must maintain a stable performance. As part of
scalability, it is also essential that the new functionalities
of the DT-ITS can expand their capabilities with the least
possible impact on existing functionalities.
3) Reliability
The DT-ITS must be highly stable to achieve reliable vir-
tual space and transport system interaction. The DT-ITS is
required to be robust enough to deal with various abnormal
situations. In case of manual operation error, illegal data,
hardware equipment failure, or other situations, the DT-ITS
must be able to handle or avoid any given error correctly. The
reliability must be guaranteed, especially when dealing with
DTs that implement time-critical transport applications. The
data modeling must also be reliable, accurately describing
the physical system’s state and predicting system operating
trends. Backing up essential data and functionality is also de-
sirable, allowing the DT-ITS to restore previous checkpoints
and critical historical points.
4) Synchronization
Strict synchronization of the DT-ITS with the transport sys-
tem is required to represent the actual state of physical enti-
ties in real time within the acceptable delay. In the same way,
the synchronization of the execution of control information
from a virtual entity to a physical entity within an acceptable
range must be addressed. This service requirement is closely
linked to the performance of the communication network and
computing platforms.
5) Security
As an essential part of current information technologies, the
DT-ITS must be secure enough to avoid and mitigate security
issues. Firstly, it must attend to the security of the data,
maintaining the confidentiality, integrity, and reliability of
the sensitive data used in the DT-ITS. Such protection must
be implemented to allow trust in unified data models and
data repositories. The DT-ITS is required to guarantee the
security of the communication and computing infrastructure
it implements. It involves software and hardware security
during data collection in the transport system and processing
of these data.
6) Privacy
The DT-ITS is required to protect the users’ private data
of the transportation system with the highest priority, com-
plying with the country’s laws based on the locations of
the transport system or its DT. In addition, the DT-ITS
must protect the privacy of the interaction between DTs and
physical entities. The information referring to the transport
system devices must be protected. The DT-ITS must be
able to handle different levels of confidential data with
varying levels of privacy, mainly operational data generated
by transport system users.
The proposed framework is our answer to RQ #3 (How
can one use DTs to address SDG regarding efficiency, safety,
and ecology in ITS?), being the first approach to accelerate
DT-ITS development. This framework even serves as a guide
to define an appropriate ontology for semantics and reference
implementations, allowing the expansion of DT-based solu-
tions in the ITS ecosystem. With a functional framework that
employs open standards, developers and practitioners of DT-
ITS solutions can focus on the applicability of current and
evolving technologies and unification strategies, developing
related business model innovations. Finally, the framework
allows, in a simpler way, to analyze scenarios from the
perspective of cross-domain interoperability, covering all the
SDGs inside and outside the transportation domain.
V. RESEARCH DIRECTIONS AND STAKEHOLDERS
The need for works on ecology pointed out in Section E
turns it into a broad research topic to investigate. However,
pressing issues still need to be addressed for the non-
functional requirements for the DT-ITS reference model
discussed above. This section provides direction on how
they could be addressed shortly and discusses new research
avenues ahead.
A. STANDARIZATION
Nowadays, the pioneering solutions using DT-ITS follow
different approaches, arbitrarily deciding their components
and relationships. Providing a functional reference frame-
work for DT-ITS has been one of the proposals of this
paper. However, the participation of standardization organi-
zations and companies in the transport sector is necessary
for a comprehensive solution, particularly non-functional
requirements. The transport vertical’s complete methodology
is domain-dependent, requiring intricate domain knowledge
to understand the DT-ITS implementation fully [144].
On the other hand, pursuing a unified and open frame-
work with a formal structure and comprehensively defined
elements of the ITS and standardized APIs would support
scalable operations of DT-ITS. In general, the standardiza-
tion process of the DT-ITS will contribute decisively to
the creation of solutions where the interactions between the
physical and virtual entities of the ITS avoid vendor lock-in,
so the accessibility to heterogeneous data for accurate cre-
ation of virtual models can be made uniformly available [48].
22 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Such frameworks must be based on a semantic alignment
of ITS by developing standardized ontologies and semantic
models for effective communication and understanding of
the modeled physical system’s concepts, relationships, and
context.
Data is essential to develop common data models that
capture the critical information and relationships within DTs,
promoting consistency, reusability, and integration across
diverse ecosystems. However, these virtual entities’ data pro-
cessing and management require more attention, especially
in transportation scenarios where critical applications may
impose additional technical requirements. As the models in
the DT-ITS change dynamically in response to alterations in
physical entities, further properties such as timestamps and
validity statements need to be standardized. Moreover, the
rise of powerful AI tools capable of extracting future value
from seemingly inert data compels organizations to reassess
their assumptions and strategies regarding data utilization
beyond immediate operational needs [8]. The distinctive
characteristics of AI impose additional demands on how
DTs handle data. Therefore, the methodologies and best
practices adopted for DTs must address these supplementary
considerations related to AI data. These special requirements
may necessitate updates to existing standards or even the
establishment of new ones.
The evaluation metrics for DT-ITS performance represent
another open research topic for developing such standards.
Given that these metrics are diverse and dependent on the
transportation domain, it is essential to focus on representa-
tive parameters that can closely reflect the ecological, safety,
and efficiency goals of ITS. To this end, extensive experi-
mentation efforts will be required, and big data analytics
will be necessary to identify relevant parameters [145]. In
this context, it is crucial to establish standardized testing
and validation procedures for DT-ITS to ensure their relia-
bility, accuracy, and performance, thereby enhancing their
credibility and trustworthiness. These procedures should
also encompass data quality assessment, model validation,
conformance testing, and performance evaluation.
The evolution of DTs can be seen as the metaverse. DTs
can be used to create the immersive virtual and persistent
online emulations of 3D virtual environments promised by
the metaverse [146]. Raising the metaverse standardization
issue here may sound far-reaching, but it should be noticed
that both the IEEE [147] and the ITU-T [148] already
have initiatives toward this end. Moreover, this ITU-T Focus
Group describes the metaverse as a potential enabler for
innovative societal problems related to SDG, and thus, stan-
dardization of the industrial metaverse for the transportation
vertical has many research avenues to be explored.
B. SCALABILITY, INFRASTRUCTURE, AND
CONNECTIVITY
The operation of ITS requires high-performance information
technology infrastructures. This is the main way to guarantee
high-fidelity two-way synchronization of the DT-ITS [149].
When considering ITS solutions empowered by DT, the
mentioned infrastructures must be able to operate, manage,
and execute intensive and computation-hungry machine and
deep learning algorithms. The high-performance graphics
processing unit considerably increases the CAPEX for im-
plementing such systems when using edge private computing
platforms. Such a situation can be circumvented with a
processing-as-a-service approach, new services that leading
companies in the public cloud market offer. However, using
solutions based solely on public clouds also implies an in-
crease in OPEX, making adopting DT-ITS unviable. Studies
are needed better to separate the workloads of the virtual
DT models and achieve a trade-off in allocating computing
[150]. In this way, further research works are needed on
the orchestration of resources in multi-tier computing. The
less computationally demanding work can be kept embedded
within the infrastructure of the ITS itself. In contrast, the
resolution of complex ML/AI algorithms would be migrated
on-demand to the cloud.
However, this orchestration effort may create other com-
plex problems related to connectivity to reach such process-
ing infrastructures. It has been recently demonstrated that
even if a model in DT-ITS is perfectly constructed, it might
still fail to predict, for instance, the trajectory of connected
vehicles. In such cloud-based algorithms, measurement and
processing imprecision and system latency may lead the
system to leave a steady convergence and enter a chaotic
region. Thus, long-term states may become unpredictable
[151].
Thus, providing reliable and low latency connectivity
between a DT and the physical system represents another
important challenge [152], especially when many sensors
need to be connected simultaneously and with real-time con-
trolling requirements for time-critical applications. Standards
related to communication technologies are not proposed
specifically for DT, but they can be reused to solve DT
problems. Considering the mobile agents of ITS, unlike con-
ventional IoT communication systems, the DT requires more
deterministic, higher broadband, better synchronization, and
other augmented transmission capabilities to enhance DT
services for diverse applications [153]. Therefore, there is
a pressing need for research works deepening B5G/6G
ultra-reliable and low-latency communications and multi-tier
computing because these two segments must come hand-
in-hand for addressing DT and metaverse issues discussed
above [154].
For the cabled segment of the infrastructure providing
connectivity, research on using time-sensitive networking
(TSN) can be an effective technique to ensure consistency
between DT elements as far as LANs are concerned [155].
This can help the vertical industry to realize the network
interconnection and data interworking of the integrated
system in wired and wireless environments and meet the
bounded business requirements of low latency, low jitter,
VOLUME , 23
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
and high reliability in various scenarios. Standards related
to communication technologies, like 5G and new WLAN
developments, are not explicitly proposed for DTs, but they
can be reused to solve DT problems. For internetwork-
ing scenarios, some issues can be dealt with deterministic
network integration along WANs [156]. Specifically, there
are comprehensive cross-platform approaches designed for
internetworking, such as OPC-UA [157], but their processing
latency is yet a research challenge to be addressed within
DT-ITS. Consequently, using these technologies efficiently
and dependably is necessary, for which formal network
analysis is a research avenue to be followed by DT-ITS in
the infrastructure design stage [158].
Even though many existing communication standards can
be reused, numerous actions are still needed for DT. It is
essential to clarify which types of equipment and parameters
should be considered for network interoperability manage-
ment, as well as how communication standards from 3GPP,
IEC, and IEEE can be extended for DT-ITS. Since real-time
performance is a key characteristic of any DT-ITS solution,
communication standards must define or quantify this real-
time aspect for various critical transportation applications.
Therefore, parameters and methods for evaluating real-time
communication should be summarized and uniformly de-
fined. Additionally, the integration and interoperability of all
these communication standards must be considered.
C. DEPENDABILITY, PRIVACY, AND SECURITY
Privacy is one of the main challenges faced by DT-ITS. For
these solutions, privacy can be compromised by the misuse
of the vast amounts of private and confidential data that DT-
ITS requires to create and develop an accurate digital replica
of physical transportation systems. Even internal operators
of DT-ITS solutions may exploit their privileges within the
system or escalate their privileges to extract information
critical to the security of the ITS that could ultimately
be used by hackers for more targeted attacks against the
system’s operation. Furthermore, in collaborative learning
scenarios [159], which are common in DT-ITS solutions,
there are potential privacy leakage risks during the DT model
aggregation process. This includes the possibility of restoring
original training samples through advanced information-
gathering techniques on edge servers [160]. Finally, in the
context of AI models applied in DT-ITS solutions, there are
potential risks of model theft when storing and delivering
the trained AI models from the cloud or edge server to the
participating entities during the cooperation between twins
[160].
In real ITS scenarios, if the DT becomes unreachable, it
could lead to hazardous situations for human life. DT-ITS, as
using real-time data exchange, is more susceptible to threats
that can make it unavailable. Threats are present in almost
all components of the DT-ITS structure, and attackers can
exploit vulnerabilities in physical systems, data structures,
software, and data communication channels. Given such a
scenario, Karaarslan and Babiker [161] analyzed the main
threats for DT-based solutions and summarized the main
countermeasures to be adopted to mitigate vulnerabilities in
the different modules of DT-ITS. DT-ITS contain critical in-
formation, allowing attackers to extract and create a mapping
of the whole system or a part of it and derive private infor-
mation or conduct patterns by analyzing databases, states,
configurations, and resources. In addition, digital assets can
make their own decisions, which can severely affect their
physical counterparts if those digital assets are deliberately
manipulated.
To preserve privacy in DT-ITS, two aspects that need
further attention are processing and analyzing data in the
DT and sharing information from simulations among the
system’s elements [162]. Firstly, while processing and an-
alyzing sensitive data in non-trusted clouds, mechanisms
for adding random noise to the raw data can be employed
to protect private information using perturbation-based and
differential privacy methods. Additionally, through encryp-
tion mechanisms, the DT-ITS can send data to computing
platforms for processing without the risk of leaking the
original information [163]. Lastly, for data sharing, data
anonymization can be applied to remove personally iden-
tifiable information, thereby maintaining the privacy of the
ITS’s users. Furthermore, federated learning enables DT-ITS
to collaboratively learn a shared model without revealing raw
data. However, more attention is needed to identify a trusted
server to maintain the global model, given the peer-to-peer
nature of DT-ITS solutions [160].
Among the main countermeasures, blockchain-based DTs
have been increasingly adopted to protect data models and
data exchange between components of the DT [164], even in
ITS solutions. While there has been tremendous fast progress
in the development of ITS, the security of such systems did
not scale up at the same rate. Blockchain emerged as an
essential tool in several sectors, including transportation. Its
decentralized, reliable, secure, transparent, and immutable
characteristics that innovate in the exchange and manage-
ment of data make it a total solution to guarantee the security
of ITS [165]. This research has proposed blockchain-enabled
DT as a service (DTaaS) for ITS to provide a secure and
reliable match between DTs and ITS. The proposal deals
with the solution for the problems of the variability of the
location and affiliation of the vehicle to the ITS. Thanks
to the features of blockchain, secure and efficient DT-ITS
service transactions, including computing, communication,
and control, can be ensured.
In ITS supported by vehicular networks, the vehicle nodes
are randomly distributed and move quickly, which leads to
the characteristics of repeated changes in the distribution of
network nodes in intelligent transportation and the uneven
distribution of the network caused by vehicular density.
Research is needed to improve DT-ITS security in these
complex transport systems. Recent work has proposed to
map the traffic situation in virtual space, using blockchain
24 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
technology to protect the vehicle and identity information on
the network [166], [167]. In addition to providing security,
DT-ITS solutions based on blockchain contribute to reducing
consensus latency in decision-making among the participants
of the transport system.
D. GOVERNANCE, BUSINESS, AND HUMAN FACTORS
The automation of the life cycle management of the DT-ITS
and the way the transportation system users manage the DT
models both need to be better developed. This management
is essential because the ITS models have been well-defined
from the technological side. Still, such evolution has not been
the same in ITS in a business context and human factors
engineering. The operation of DT-ITS solutions requires the
involvement of many parties and users (e.g., government,
transport groups, automotive companies, road users, etc.), so
the existence of well-defined business models will facilitate
access to DT-ITS resources, enabling their efficient manage-
ment in multi-user and multi-operator scenarios. Moreover,
a proper business layer might need to be added to our
proposal in Fig. 7, like in the reference model in RAMI
4.0 [27], to automate communication across businesses. In
this sense, it must consider the practitioner experiences and
convenience through better-suited products and satisfaction,
directly influenced by the fidelity of the results obtained
through the DT-ITS [168]. In addition, the relevant legal and
ethical issues in a governance model for this interdisciplinary
and multi-stakeholder DT-ITS are yet to be fully understood.
Misplacing roles that humans and DTs play may result
in significant problems not only for the early adopters but
also for the viability of DT-ITS solutions themselves due
to irreparable reputation harm caused by unrealistic expec-
tations and strategic misalignments [169]. Therefore, in the
design phase, human factors must be considered. For DT-ITS
to deserve our reliance on it, engineers and managers alike
must trust frameworks, methods, and tools that will replace
legacy systems. Experimental evidence and case studies as
benchmark problems should be provided to gradually build
confidence that DTs are fit for purpose [170]. In this context,
determining when to enable a DT-ITS to be autonomous and
when to rely on humans to make decisions is another relevant
question.
E. STAKEHOLDERS’ PERSPECTIVE
Meeting stakeholders’ often conflicting expectations is
paramount for the success of any digital infrastructure
project, including DT-ITS. Identifying key stakeholders, in-
cluding government agencies, private sector partners, and
the public, and outlining how they will be involved in the
development and implementation process is vital for DT-
ITS [171]. Primary stakeholders, including business owners,
managers, and developers, are easier to identify because
they are directly involved with the development process
and implementation of the DT-ITS. Nonetheless, based on
the conclusions learned from this survey, identifying and
collaborating with secondary stakeholders, including those
who indirectly affect and/or are affected by the DT-ITS,
such as local authorities or local communities, is not com-
mon practice except in very specific cases such as those
mentioned in the case studies. This problem is becoming
a challenge to achieve stakeholder engagement in adopting
DT-ITS [172].
Stakeholders’ perspectives range from technical require-
ments for data and functionality to business considerations
such as revenue generation and market opportunities, as
well as organizational issues like collaboration and com-
munication [173]. Although standard-setting bodies are not
considered stakeholders, they employ mechanisms for stake-
holder participation in the standard-setting process for the
successful implementation of DT-ITS [48]. By developing
and promoting standards, these organizations facilitate data
integration, system interoperability, coordination among ac-
tors, and trust in data, contributing to value creation in
DT-ITS. Governments and companies collaborate to create
standards that adhere to the guidelines set by standardization
organizations and promote them.
For DT-ITS, we suggest DT-ITS researchers consider the
following stakeholders: citizens as end-users of the trans-
portation system, IT enterprises as developers of DT-ITS,
and government or local authorities as policymakers [174].
Citizens play an important role in the development and im-
plementation of DT-ITS, although their influence and type of
participation may vary throughout the process, making their
role dynamic and multifaceted [174]. As end-users of DT-
ITS, whether as drivers or vulnerable users of the ITS, they
utilize DT-ITS facilities and applications to access the ITS
and take advantage of available services, making them the
primary consumers of data generated by the operation of DT-
ITS. However, they are not limited to being data consumers.
The users also become data producers by interacting with
DT-ITS solutions. Their feedback and active participation
are essential for other stakeholders to implement continuous
improvements, ensure the sustainability of DT-ITS, and
foster the development of new solutions. To accelerate the
adoption of DT-based solutions, citizens are encouraged to
adopt a ”smart” mindset, which involves the conscious and
proactive use of available technologies and alignment with
sustainability principles [175].
Private companies play another fundamental role in de-
veloping and implementing DT-ITS, primarily responsible
for constructing the necessary data infrastructure through en-
abling technologies. Additionally, they design and implement
DT-based user-centered applications for various transport
scenarios, aiming to connect users with the administrative
authorities of the transport system [105]. In this way, private
companies are considered key stakeholders in creating value
for DT-ITS, as they collaborate with government and local
authorities to implement solutions that address the users’
needs with a focus on sustainability objectives. They must
work alongside other stakeholders in the implementation
VOLUME , 25
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
of pilot projects, which serve as demonstrations of the
technological capabilities of DT-ITS, and actively participate
in the standardization processes of these solutions. In short,
private companies play a dynamic role in creating a DT-
ITS ecosystem that adapts to the changing needs of society,
acting as intermediaries between the government and users to
improve existing services and develop new solutions [175].
Governments and local authorities have enacted policies
related to DT-ITS and introduced a myriad of innovative
projects, such as living labs [11], [12], [131]. Governments
play a leading role in coordinating various DT-ITS initiatives
and standardizing them at the national level-areas that are
critical, ensuring that data from different systems can be
effectively integrated and used for transport sustainability.
They are responsible for accrediting and investing in DT-
based transport solutions, setting the legal and policy frame-
work for adopting and expanding DT-ITS initiatives [176].
At the same time, governments and local authorities must en-
act policies for data protection and cybersecurity, which other
stakeholders must strictly enforce. Additionally, governments
encourage collaborations that facilitate the exchange of data
and the creation of innovative solutions based on DT by
establishing public-private partnerships among stakeholders.
Finally, the participation of governments through financial
investments is an essential priority to achieve efficient and
sustainable ITS.
VI. CONCLUSIONS
Digital twins can be pivotal in revolutionizing sustainable
ITS. By creating virtual replicas of physical assets, DTs offer
a comprehensive understanding of infrastructure, vehicles,
and operational processes, with the potential to optimize
energy consumption, traffic flow, and resource allocation.
Driven by the SDG from the UN, we undertook an ex-
haustive survey in this paper to encapsulate the core aspects
of DTs that could trigger sustainable applications in ITS
aimed at fulfilling diverse (and often conflicting) objectives.
We formulated and answered three fundamental research
questions, paving the way for adopting DTs in ITS. From
the key characteristics in the surveyed works using DT for
ITS, we introduced a concise 4-layer reference framework
for creating DT-ITS. We deliberated on its functional and
non-functional prerequisites.
The framework is the first step towards expediting the
development of DTs for ITS. The framework acts as a
guide to establish an appropriate ontology for semantics
and reference implementations, facilitating the proliferation
of DT-based solutions in the ITS ecosystem. Moreover, we
underscored several open challenges and future research
avenues for developing DT solutions in ITS, indicating
potential opportunities for further exploration and improve-
ment.
ACKNOWLEDGMENT
V´
ıctor M. G. Mart´
ınez, Divanilson R. Campelo, and Mois´
es
R. N. Ribeiro, as the authors, collectively contributed to the
conception of this work. The author would like to thank
Salim S. Musi for his continuous support and valuable
discussion throughout the research.
REFERENCES
[1] J. Sussman, Introduction to Transportation Systems. Artech House,
2000.
[2] (2015) Transforming our world: the 2030 agenda for sustainable
development. [Online]. Available: https://sdgs.un.org/2030agenda
[3] L. Deka, S. Khan, M. Chowdhury, and N. Ayres, Transportation
Cyber-Physical System and its importance for future mobility, 01
2018, pp. 1–20.
[4] F. Wang and J. J. Zhang, “Transportation 5.0 in cpss: Towards
acp-based society-centered intelligent transportation,” in 2017 IEEE
20th International Conference on Intelligent Transportation Systems
(ITSC), 10 2017, pp. 762–767.
[5] L. Yongfu, S. Dihua, L. Weining, and Z. Xuebo, “A service-oriented
architecture for the transportation cyber-physical systems,” in Pro-
ceedings of the 31st Chinese Control Conference, 07 2012, pp. 7674–
7678.
[6] V. Kant, “Cyber-physical systems as sociotechnical systems: a view
towards human–technology interaction,” Cyber-Physical Systems,
vol. 2, no. 1-4, pp. 75–109, 2016.
[7] Q. Qi, F. Tao, T. Hu, N. Anwer, A. Liu, Y. Wei, L. Wang, and
A. Nee, “Enabling technologies and tools for digital twin,” Journal
of Manufacturing Systems, vol. 58, pp. 3–21, 2021.
[8] M. M. Rathore, S. A. Shah, D. Shukla, E. Bentafat, and S. Bakiras,
“The role of ai, machine learning, and big data in digital twinning:
A systematic literature review, challenges, and opportunities,” IEEE
Access, vol. 9, pp. 32 030–32 052, 2021.
[9] T. H. Luan, R. Liu, L. Gao, R. Li, and H. Zhou, “The paradigm of
digital twin communications,” 2021.
[10] K. Wang, Y. Wang, Y. Li, X. Fan, S. Xiao, and L. Hu, “A review
of the technology standards for enabling digital twin,” Digital Twin,
vol. 2, no. 4, 2022.
[11] H. Xu, A. Berres, S. B. Yoginath, H. Sorensen, P. J. Nugent,
J. Severino, S. A. Tennille, A. Moore, W. Jones, and J. Sanyal, “Smart
mobility in the cloud: Enabling real-time situational awareness and
cyber-physical control through a digital twin for traffic,” IEEE
Transactions on Intelligent Transportation Systems, vol. 24, no. 3,
pp. 3145–3156, 2023.
[12] L. Raes, P. Michiels, T. Adolphi, C. Tampere, A. Dalianis,
S. McAleer, and P. Kogut, “Duet: A framework for building in-
teroperable and trusted digital twins of smart cities,” IEEE Internet
Computing, vol. 26, no. 3, pp. 43–50, 2022.
[13] A. D. B. (ADB), “Building greener, resilient transport infrastructure:
Innovating with artificial intelligence and digital twins in road design
in uzbekistan,” Tech. Rep., 2023.
[14] Z. Hu, S. Lou, Y. Xing, X. Wang, D. Cao, and C. Lv, “Review
and perspectives on driver digital twin and its enabling technologies
for intelligent vehicles,” IEEE Transactions on Intelligent Vehicles,
vol. 7, no. 3, pp. 417–440, 2022.
[15] B. R. Barricelli, E. Casiraghi, and D. Fogli, “A survey on digital twin:
Definitions, characteristics, applications, and design implications,”
IEEE Access, vol. 7, pp. 167 653–167 671, 2019.
[16] A. Rasheed, O. San, and T. Kvamsdal, “Digital twin: Values, chal-
lenges and enablers from a modeling perspective,” IEEE Access,
vol. 8, pp. 21 980–22 012, 2020.
[17] R. Minerva, G. M. Lee, and N. Crespi, “Digital twin in the iot context:
A survey on technical features, scenarios, and architectural models,”
Proceedings of the IEEE, vol. 108, no. 10, pp. 1785–1824, 2020.
[18] D. Jones, C. Snider, A. Nassehi, J. Yon, and B. Hicks, “Characterising
the digital twin: A systematic literature review,” CIRP Journal of
Manufacturing Science and Technology, vol. 29, pp. 36–52, 2020.
[19] C. Semeraro, M. Lezoche, H. Panetto, and M. Dassisti, “Digital twin
paradigm: A systematic literature review,” Computers in Industry,
vol. 130, 9 2021.
26 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
[20] M. Liu, S. Fang, H. Dong, and C. Xu, “Review of digital twin
about concepts, technologies, and industrial applications,” Journal
of Manufacturing Systems, vol. 58, pp. 346–361, 2021, digital Twin
towards Smart Manufacturing and Industry 4.0.
[21] M. Singh, E. Fuenmayor, E. P. Hinchy, Y. Qiao, N. Murray, and
D. Devine, “Digital twin: Origin to future,” Applied System Innova-
tion, vol. 4, no. 2, 2021.
[22] Y. K. Liu, S. K. Ong, and A. Y. Nee, “State-of-the-art survey on
digital twin implementations,” Advances in Manufacturing, 2022.
[23] F. Tao, H. Zhang, A. Liu, and A. Y. C. Nee, “Digital twin in indus-
try: State-of-the-art,” IEEE Transactions on Industrial Informatics,
vol. 15, no. 4, pp. 2405–2415, 2019.
[24] T. Y. Melesse, V. Di Pasquale, and S. Riemma, “Digital twin models
in industrial operations: State-of-the-art and future research direc-
tions,” IET Collaborative Intelligent Manufacturing, vol. 3, no. 1,
pp. 37–47, 2021.
[25] M. Vukovi´
c, D. Mazzei, S. Chessa, and G. Fantoni, “Digital twins
in industrial iot: a survey of the state of the art and of relevant stan-
dards,” in 2021 IEEE International Conference on Communications
Workshops (ICC Workshops), 2021, pp. 1–6.
[26] M. Perno, L. Hvam, and A. Haug, “Implementation of digital twins
in the process industry: A systematic literature review of enablers
and barriers,” Computers in Industry, vol. 134, p. 103558, 2022.
[27] M. Lindner, L. Bank, J. Schilp, and M. Weigold, “Digital twins in
manufacturing: A rami 4.0 compliant concept,” Sci, vol. 5, no. 4,
2023.
[28] C. Coupry, S. Noblecourt, P. Richard, D. Baudry, and D. Bigaud,
“Bim-based digital twin and xr devices to improve maintenance
procedures in smart buildings: A literature review,” Applied Sciences,
vol. 11, no. 15, 2021.
[29] F. Jiang, L. Ma, T. Broyd, and K. Chen, “Digital twin and its
implementations in the civil engineering sector,” Automation in
Construction, vol. 130, p. 103838, 2021.
[30] M. Callcut, J.-P. Cerceau Agliozzo, L. Varga, and L. McMillan,
“Digital twins in civil infrastructure systems,” Sustainability, vol. 13,
no. 20, 2021.
[31] D. G. Broo and J. Schooling, “Digital twins in infrastructure: def-
initions, current practices, challenges and strategies,” International
Journal of Construction Management, vol. 0, no. 0, pp. 1–10, 2021.
[32] D.-G. J. Opoku, S. Perera, R. Osei-Kyei, and M. Rashidi, “Digital
twin application in the construction industry: A literature review,”
Journal of Building Engineering, vol. 40, p. 102726, 2021.
[33] D.-G. J. Opoku, S. Perera, R. Osei-Kyei, M. Rashidi, T. Famakinwa,
and K. Bamdad, “Drivers for digital twin adoption in the construction
industry: A systematic literature review,” Buildings, vol. 12, no. 2,
2022.
[34] T. R. Sreedevi and M. Santosh Kumar, “Digital twin in smart
farming: A categorical literature review and exploring possibilities
in hydroponics,” in 2020 Advanced Computing and Communication
Technologies for High Performance Applications (ACCTHPA), 2020,
pp. 120–124.
[35] C. Pylianidis, S. Osinga, and I. N. Athanasiadis, “Introducing digital
twins to agriculture,” Computers and Electronics in Agriculture, vol.
184, p. 105942, 2021.
[36] C. Verdouw, B. Tekinerdogan, A. Beulens, and S. Wolfert, “Digital
twins in smart farming,” Agricultural Systems, vol. 189, p. 103046,
2021.
[37] S. Nguyen, M. Abdelhakim, and R. Kerestes, “Survey paper of digital
twins and their integration into electric power systems,” in 2021 IEEE
Power Energy Society General Meeting (PESGM), 2021, pp. 01–05.
[38] X. Zhang, J. Shen, P. K. Saini, M. Lovati, M. Han, P. Huang, and
Z. Huang, “Digital twin for accelerating sustainability in positive
energy district: A review of simulation tools and applications,”
Frontiers in Sustainable Cities, vol. 3, 2021.
[39] A. E. Onile, R. Machlev, E. Petlenkov, Y. Levron, and J. Belikov,
“Uses of the digital twins concept for energy services, intelligent
recommendation systems, and demand side management: A review,”
Energy Reports, vol. 7, pp. 997–1015, 2021.
[40] A. Aghazadeh Ardebili, A. Longo, and A. Ficarella, “Digital twins
bonds society with cyber-physical energy systems: a literature re-
view,” in 2021 IEEE International Conferences on Internet of Things
(iThings) and IEEE Green Computing Communications (GreenCom)
and IEEE Cyber, Physical Social Computing (CPSCom) and IEEE
Smart Data (SmartData) and IEEE Congress on Cybermatics (Cy-
bermatics), 2021, pp. 284–289.
[41] G. Ahmadi-Assalemi, H. Al-Khateeb, C. Maple, G. Epiphaniou, Z. A.
Alhaboby, S. Alkaabi, and D. Alhaboby, Digital Twins for Precision
Healthcare. Cham: Springer International Publishing, 2020, pp.
133–158.
[42] P.-h. Huang, K.-h. Kim, and M. Schermer, “Ethical issues of digital
twins for personalized health care service: Preliminary mapping
study,” J Med Internet Res, vol. 24, no. 1, p. e33081, Jan 2022.
[43] G. Mylonas, A. Kalogeras, G. Kalogeras, C. Anagnostopoulos,
C. Alexakos, and L. Mu˜
noz, “Digital twins from smart manufacturing
to smart cities: A survey,” IEEE Access, vol. 9, pp. 143 222–143249,
2021.
[44] S. P. Ramu, P. Boopalan, Q.-V. Pham, P. K. R. Maddikunta, T. Huynh-
The, M. Alazab, T. T. Nguyen, and T. R. Gadekallu, “Federated
learning enabled digital twins for smart cities: Concepts, recent
advances, and future directions,” Sustainable Cities and Society,
vol. 79, p. 103663, 2022.
[45] L. Chen, Y. Li, C. Huang, B. Li, Y. Xing, D. Tian, L. Li, Z. Hu,
X. Na, Z. Li, S. Teng, C. Lv, J. Wang, D. Cao, N. Zheng, and F.-
Y. Wang, “Milestones in autonomous driving and intelligent vehicles:
Survey of surveys,” IEEE Transactions on Intelligent Vehicles, vol. 8,
no. 2, pp. 1046–1056, 2023.
[46] M. S. Irfan, S. Dasgupta, and M. Rahman, “Toward transportation
digital twin systems for traffic safety and mobility: A review,” IEEE
Internet of Things Journal, vol. 11, no. 14, pp. 24 581–24 603, 2024.
[47] C. Ge and S. Qin, “Digital twin intelligent transportation system (dt-
its)—a systematic review,” IET Intelligent Transport Systems, vol. 18,
no. 12, pp. 2325–2358, 2024.
[48] ITU-T, “Requirements and capability framework of digital twin for
intelligent transport system,” International Telecommunication Union,
Recommendation Y.dt-ITS, 2021.
[49] X. Liao, Z. Wang, X. Zhao, K. Han, P. Tiwari, M. J. Barth, and
G. Wu, “Cooperative ramp merging design and field implementation:
A digital twin approach based on vehicle-to-cloud communication,”
IEEE Transactions on Intelligent Transportation Systems, vol. 23,
no. 5, pp. 4490–4500, 2022.
[50] F. Jiang, L. Ma, T. Broyd, W. Chen, and H. Luo, “Digital twin enabled
sustainable urban road planning,” Sustainable Cities and Society,
vol. 78, p. 103645, 2022.
[51] C. Stadler, F. Montanari, W. Baron, C. Sippl, and A. Djanatliev,
“A credibility assessment approach for scenario-based virtual testing
of automated driving functions,” IEEE Open Journal of Intelligent
Transportation Systems, vol. 3, pp. 45–60, 2022.
[52] Z. Liu, X. Mao, H. Zhang, and Y. Li, “Driving in the virtual: Towards
large-scale software-defined transportation scene,” IEEE Journal of
Radio Frequency Identification, pp. 1–1, 2022.
[53] C. Sun, Z. Deng, W. Chu, S. Li, and D. Cao, “Acclimatizing the
operational design domain for autonomous driving systems,” IEEE
Intelligent Transportation Systems Magazine, vol. 14, no. 2, pp. 10–
24, 2022.
[54] B. Yu, C. Chen, J. Tang, S. Liu, and J.-L. Gaudiot, “Autonomous
vehicles digital twin: A practical paradigm for autonomous driving
system development,” Computer, vol. 55, no. 9, pp. 26–34, 2022.
[55] D. An, N. H. Kim, and J.-H. Choi, “Practical options for selecting
data-driven or physics-based prognostics algorithms with reviews,”
Reliability Engineering & System Safety, vol. 133, pp. 223–236, 2015.
[56] Y. Zhang and H. Zhang, “Urban digital twins: Decision-making
models for transportation network simulation,” in Proceedings of
the 2022 International Conference on Computational Infrastructure
and Urban Planning, ser. CIUP ’22. Association for Computing
Machinery, 2022, p. 18–21.
[57] G. Papyshev and M. Yarime, “Exploring city digital twins as policy
tools: A task-based approach to generating synthetic data on urban
mobility,” Data & Policy, vol. 3, p. e16, 2021.
[58] H. Naing, W. Cai, N. Hu, T. Wu, and L. Yu, “Data-driven microscopic
traffic modelling and simulation using dynamic lstm,” ser. SIGSIM-
PADS ’21. Association for Computing Machinery, 2021, p. 1–12.
[59] Q. Han, L. Qi, L. Ye, H. Zhang, L. Zeng, and D. Zhang, “On-road test
scenarios extraction and re-construction approach for c-v2x digital
twin test,” in 2022 IEEE 25th International Conference on Intelligent
Transportation Systems (ITSC), 2022, pp. 3269–3275.
[60] P. Brunner, F. Denk, W. Huber, and R. Kates, “Virtual safety
performance assessment for automated driving in complex urban
VOLUME , 27
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
traffic scenarios,” in 2019 IEEE Intelligent Transportation Systems
Conference (ITSC), 2019, pp. 679–685.
[61] Z. Szalay, M. Szalai, B. T´
oth, T. Tettamanti, and V. Tihanyi, “Proof of
concept for scenario-in-the-loop (scil) testing for autonomous vehicle
technology,” in 2019 IEEE International Conference on Connected
Vehicles and Expo (ICCVE), 2019, pp. 1–5.
[62] B. Varga, T. Tettamanti, and Z. Szalay, “System architecture for
scenario-in-the-loop automotive testing,” Transport and Telecommu-
nication Journal, vol. 22, no. 2, pp. 141–151, 2021.
[63] T. Tettamanti, M. Szalai, S. Vass, and V. Tihanyi, “Vehicle-in-the-loop
test environment for autonomous driving with microscopic traffic
simulation,” in 2018 IEEE International Conference on Vehicular
Electronics and Safety (ICVES), 2018, pp. 1–6.
[64] C. Schwarz and Z. Wang, “The role of digital twins in connected
and automated vehicles,” IEEE Intelligent Transportation Systems
Magazine, vol. 14, no. 6, pp. 41–51, 2022.
[65] D. Brecht, N. Gehrke, T. Kerbl, N. Krauss, D. Majstorovi´
c, F. Pfab,
M.-M. Wolf, and F. Diermeyer, “Evaluation of teleoperation concepts
to solve automated vehicle disengagements,” IEEE Open Journal of
Intelligent Transportation Systems, vol. 5, pp. 629–641, 2024.
[66] T. Xiao, S. Chen, K. Zhu, R. Huang, A. Li, N. Zheng, and J. Xin, “In-
jection simulation: an efficient validation framework for autonomous
driving system,” in 2022 IEEE 25th International Conference on
Intelligent Transportation Systems (ITSC), 2022, pp. 472–478.
[67] M. U. Shoukat, S. Yu, S. Shi, Y. Li, and J. Yu, “Evaluate the
connected autonomous vehicles infrastructure using digital twin
model based on cyber-physical combination of intelligent network,”
in 2021 5th CAA International Conference on Vehicular Control and
Intelligence (CVCI), 2021, pp. 1–6.
[68] M. Szalai, B. Varga, T. Tettamanti, and V. Tihanyi, “Mixed reality
test environment for autonomous cars using unity 3d and sumo,” in
2020 IEEE 18th World Symposium on Applied Machine Intelligence
and Informatics (SAMI), 2020, pp. 73–78.
[69] L. Shuguang, L. Zhonglin, W. Wenbo, Z. Yang, H. Jierui, and
C. Hong, “Vehicle-in-the-loop intelligent connected vehicle simula-
tion system based on vehicle-road-cloud collaboration,” in 2022 IEEE
25th International Conference on Intelligent Transportation Systems
(ITSC), 2022, pp. 1711–1716.
[70] Z. Fan, X. Yang, W. Yuan, R. Jiang, Q. Chen, X. Song, and
R. Shibasaki, “Online trajectory prediction for metropolitan scale
mobility digital twin,” in Proceedings of the 30th International
Conference on Advances in Geographic Information Systems,
ser. SIGSPATIAL ’22. New York, NY, USA: Association
for Computing Machinery, 2022. [Online]. Available: https:
//doi.org/10.1145/3557915.3561040
[71] K. Ling, “Digital twinning from vehicle usage statistics for customer-
centric automotive systems engineering,” IEEE Open Journal of
Intelligent Transportation Systems, vol. 4, pp. 966–978, 2023.
[72] Z. Wang, X. Liao, X. Zhao, K. Han, P. Tiwari, M. J. Barth, and
G. Wu, “A digital twin paradigm: Vehicle-to-cloud based advanced
driver assistance systems,” in 2020 IEEE 91st Vehicular Technology
Conference (VTC2020-Spring), 2020, pp. 1–6.
[73] Y. Wang, Z. Wang, K. Han, P. Tiwari, and D. B. Work, “Gaussian
process-based personalized adaptive cruise control,” IEEE Trans-
actions on Intelligent Transportation Systems, vol. 23, no. 11, pp.
21 178–21 189, 2022.
[74] K. Kuru, “Conceptualisation of human-on-the-loop haptic teleopera-
tion with fully autonomous self-driving vehicles in the urban envi-
ronment,” IEEE Open Journal of Intelligent Transportation Systems,
vol. 2, pp. 448–469, 2021.
[75] Z. Wang, K. Han, and P. Tiwari, “Digital twin simulation of con-
nected and automated vehicles with the unity game engine,” in 2021
IEEE 1st International Conference on Digital Twins and Parallel
Intelligence (DTPI), 2021, pp. 1–4.
[76] R. Tang and Z. Jiang, “Driver’s perception model in driving assist,”
in 2020 IEEE 20th International Conference on Software Quality,
Reliability and Security Companion (QRS-C), 2020, pp. 237–240.
[77] Y. Liu, Z. Wang, K. Han, Z. Shou, P. Tiwari, and J. H. L. Hansen,
“Sensor fusion of camera and cloud digital twin information for
intelligent vehicles,” in 2020 IEEE Intelligent Vehicles Symposium
(IV), 2020, pp. 182–187.
[78] Y. Liu, Z. Wang, K. Han, Z. Shou, P. Tiwari, and J. H. L. Hansen,
“Vision-cloud data fusion for adas: A lane change prediction case
study,” IEEE Transactions on Intelligent Vehicles, vol. 7, no. 2, pp.
210–220, 2022.
[79] D. Li, J. Zhang, B. Xiao, B. Tang, and Z. Hu, “Vehicle crash mitiga-
tion strategy in unavoidable collision scenarios: focusing on motion
planning by considering a generalized crash severity model,” Journal
of the Brazilian Society of Mechanical Sciences and Engineering,
vol. 44, no. 12, p. 581, Nov 2022.
[80] B. Wang, C. Zhang, M. Zhang, C. Liu, Z. Xie, and H. Zhang, “Digital
twin analysis for driving risks based on virtual physical simulation
technology,” IEEE Journal of Radio Frequency Identification, pp. 1–
1, 2022.
[81] C. Yang, J. Dong, Q. Xu, M. Cai, H. Qin, J. Wang, and K. Li, “Multi-
vehicle experiment platform: A digital twin realization method,” in
2022 IEEE/SICE International Symposium on System Integration
(SII), 2022, pp. 705–711.
[82] Z. Lv, D. Chen, and M. S. Hossain, “Traffic safety detection system
by digital twins and virtual reality technology,” in 2022 IEEE Inter-
national Instrumentation and Measurement Technology Conference
(I2MTC), 2022, pp. 1–6.
[83] Z. Lv, J. Guo, A. K. Singh, and H. Lv, “Digital twins based
vr simulation for accident prevention of intelligent vehicle,” IEEE
Transactions on Vehicular Technology, vol. 71, no. 4, pp. 3414–3428,
2022.
[84] J. Lu, S. Yang, R. Shi, Y. Chen, and Y. Cao, “Safety co-pilot: A
system for autonomous vehicle to make decision safer and smarter,”
in 2022 IEEE 25th International Conference on Intelligent Trans-
portation Systems (ITSC), 2022, pp. 1809–1816.
[85] O. Veledar, V. Damjanovic-Behrendt, and G. Macher, “Digital twins
for dependability improvement of autonomous driving,” in Systems,
Software and Services Process Improvement, A. Walker, R. V.
O’Connor, and R. Messnarz, Eds. Cham: Springer International
Publishing, 2019, pp. 415–426.
[86] S. Almeaibed, S. Al-Rubaye, A. Tsourdos, and N. P. Avdelidis,
“Digital twin analysis to promote safety and security in autonomous
vehicles,” IEEE Communications Standards Magazine, vol. 5, no. 1,
pp. 40–46, 2021.
[87] S. A. P. Kumar, R. Madhumathi, P. R. Chelliah, L. Tao, and S. Wang,
“A novel digital twin-centric approach for driver intention prediction
and traffic congestion avoidance,” Journal of Reliable Intelligent
Environments, vol. 4, pp. 199–209, 2018.
[88] X. Chen, E. Kang, S. Shiraishi, V. M. Preciado, and Z. Jiang,
“Digital behavioral twins for safe connected cars,” in Proceedings
of the 21th ACM/IEEE International Conference on Model Driven
Engineering Languages and Systems, ser. MODELS ’18. New York,
NY, USA: Association for Computing Machinery, 2018, p. 144–153.
[Online]. Available: https://doi.org/10.1145/3239372.3239401
[89] Z. Wang, K. Han, and P. Tiwari, “Digital twin-assisted cooperative
driving at non-signalized intersections,” IEEE Transactions on Intel-
ligent Vehicles, vol. 7, no. 2, pp. 198–209, 2022.
[90] Y. Hui, X. Ma, Z. Su, N. Cheng, Z. Yin, T. H. Luan, and Y. Chen,
“Collaboration as a service: Digital-twin-enabled collaborative and
distributed autonomous driving,” IEEE Internet of Things Journal,
vol. 9, no. 19, pp. 18 607–18 619, 2022.
[91] Z. Xu, T. Jiang, and N. Zheng, “Developing and analyzing eco-
driving strategies for on-road emission reduction in urban transport
systems - a vr-enabled digital-twin approach,” Chemosphere, vol.
305, p. 135372, 2022.
[92] Y. Liang, Z. Yin, L. Nie, and Y. Ba, “Shared steering control with
predictive risk field enabled by digital twin,” IEEE Transactions on
Intelligent Vehicles, vol. 8, no. 5, pp. 3256–3269, 2023.
[93] K. Yang, C. A. Haddad, G. Yannis, and C. Antoniou, “Classification
and evaluation of driving behavior safety levels: A driving simulation
study,” IEEE Open Journal of Intelligent Transportation Systems,
vol. 3, pp. 111–125, 2022.
[94] H. Neuschmied, F. Krebs, S. Ladst¨
atter, E. Eder, M. R. Berrazouane,
and G. Thallinger, “A toolchain for extracting and visualising road
traffic data,” in Proceedings of the 19th International Conference on
Content-Based Multimedia Indexing, ser. CBMI ’22. New York, NY,
USA: Association for Computing Machinery, 2022, p. 205–208.
[95] S. Wang, F. Zhang, and T. Qin, “Research on the construction of
highway traffic digital twin system based on 3d gis technology,”
Journal of Physics: Conference Series, vol. 1802, no. 4, p. 042045,
mar 2021.
28 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
[96] Y. Guo, K. Zou, S. Chen, F. Yuan, and F. Yu, “3d digital twin of
intelligent transportation system based on road-side sensing,” Journal
of Physics: Conference Series, vol. 2083, no. 3, p. 032022, nov 2021.
[97] O. E. Marai, T. Taleb, and J. Song, “Roads infrastructure digital twin:
A step toward smarter cities realization,” IEEE Network, vol. 35,
no. 2, pp. 136–143, 2021.
[98] G. Mao, Y. Hui, X. Ren, C. Li, and Y. Shao, “The internet of
things for smart roads: A road map from present to future road
infrastructure,” IEEE Intelligent Transportation Systems Magazine,
vol. 14, no. 6, pp. 66–76, 2022.
[99] C. Liu, C. Zhang, B. Wang, Z. Tang, and Z. Xie, “Digital twin of
highway entrances and exits: a traffic risk identification method,”
IEEE Journal of Radio Frequency Identification, pp. 1–1, 2022.
[100] X. Ji, W. Yue, C. Li, Y. Chen, N. Xue, and Z. Sha, “Digital twin
empowered model free prediction of accident-induced congestion
in urban road networks,” in 2022 IEEE 95th Vehicular Technology
Conference: (VTC2022-Spring), 2022, pp. 1–6.
[101] H. Fu, T. Zhao, Y. Chen, Y. Yao, and J. Leng, “Framework and
operation of digital twin smart freeway,” IET Intelligent Transport
Systems, vol. n/a, no. n/a, 2022.
[102] K. Kuˇ
si´
c, R. Schumann, and E. Ivanjko, “Building a motorway digital
twin in sumo: Real-time simulation of continuous data stream from
traffic counters,” in 2022 International Symposium ELMAR, 2022,
pp. 71–76.
[103] V. Mart´
ınez, M. Ribeiro, and D. Campelo, “Intelligent road intersec-
tions: A case for digital twins,” in Anais do III Workshop Brasileiro
de Cidades Inteligentes. Porto Alegre, RS, Brasil: SBC, 2022, pp.
151–158.
[104] C. Zhang, L. Zhu, and C. Xu, “Bsdp: Blockchain-based smart parking
for digital-twin empowered vehicular sensing networks with privacy
protection,” IEEE Transactions on Industrial Informatics, vol. 19,
no. 5, pp. 7237–7246, 2023.
[105] E. Thonhofer, S. Sigl, M. Fischer, F. Heuer, A. Kuhn, J. Erhart,
M. Harrer, and W. Schildorfer, “Infrastructure-based digital twins for
cooperative, connected, automated driving and smart road services,”
IEEE Open Journal of Intelligent Transportation Systems, vol. 4, pp.
311–324, 2023.
[106] O. Zheng, M. Abdel-Aty, L. Yue, A. Abdelraouf, Z. Wang, and
N. Mahmoud, “Citysim: A drone-based vehicle trajectory dataset for
safety-oriented research and digital twins,” Transportation Research
Record, vol. 2678, no. 4, pp. 606–621, 2024.
[107] X. Chen, M. Hou, T. Tang, A. Kaur, and F. Xia, “Digital twin mobility
profiling: A spatio-temporal graph learning approach,” in 2021 IEEE
23rd Int Conf on High Performance Computing & Communications;
7th Int Conf on Data Science & Systems; 19th Int Conf on Smart City;
7th Int Conf on Dependability in Sensor, Cloud & Big Data Systems
& Application (HPCC/DSS/SmartCity/DependSys), 2021, pp. 1178–
1187.
[108] Y. Shi, Z. Liu, Z. Wang, J. Ye, W. Tong, and Z. Liu, “An integrated
traffic and vehicle co-simulation testing framework for connected
and autonomous vehicles,” IEEE Intelligent Transportation Systems
Magazine, vol. 14, no. 6, pp. 26–40, 2022.
[109] T. Zhang, X. Liu, Z. Luo, F. Dong, and Y. Jiang, “Time series be-
havior modeling with digital twin for internet of vehicles,” EURASIP
Journal on Wireless Communications and Networking, vol. 2019,
no. 1, p. 271, Dec 2019.
[110] A. J. Saroj, S. Roy, A. Guin, and M. Hunter, “Development of a
connected corridor real-time data-driven traffic digital twin simulation
model,” Journal of Transportation Engineering, Part A: Systems, vol.
147, no. 12, p. 04021096, 2021.
[111] Y. Hui, Q. Wang, N. Cheng, R. Chen, X. Xiao, and T. H. Luan, “Time
or reward: Digital-twin enabled personalized vehicle path planning,”
in 2021 IEEE Global Communications Conference (GLOBECOM),
2021, pp. 1–6.
[112] A. I. Zhukov and D. G. Moroz, “Simulation modeling of a bus route,”
in 2021 Intelligent Technologies and Electronic Devices in Vehicle
and Road Transport Complex (TIRVED), 2021, pp. 1–5.
[113] S. Deng, J. Zhong, S. Chen, and Z. He, “Digital twin modeling
for demand responsive transit,” in 2021 IEEE 1st International
Conference on Digital Twins and Parallel Intelligence (DTPI), 2021,
pp. 410–413.
[114] L. Li, Z. Hu, and X. Yang, “Intelligent analysis of abnormal vehicle
behavior based on a digital twin,” Journal of Shanghai Jiaotong
University (Science), vol. 26, no. 5, pp. 587–597, Oct 2021.
[115] Z. Wang, R. Gupta, K. Han, H. Wang, A. Ganlath, N. Ammar, and
P. Tiwari, “Mobility digital twin: Concept, architecture, case study,
and future challenges,” IEEE Internet of Things Journal, vol. 9,
no. 18, pp. 17 452–17 467, 2022.
[116] C. Hu, W. Fan, E. Zeng, Z. Hang, F. Wang, L. Qi, and M. Z. A.
Bhuiyan, “Digital twin-assisted real-time traffic data prediction
method for 5g-enabled internet of vehicles,” IEEE Transactions on
Industrial Informatics, vol. 18, no. 4, pp. 2811–2819, 2022.
[117] S. Wang, F. Zhang, and T. Qin, “Urban traffic optimal solution
based on digital twin system technology simulation,” in Proceedings
of the 14th International Conference on Computer Modeling and
Simulation, ser. ICCMS ’22, 2022, p. 83–87.
[118] U. Ritzinger, J. Puchinger, C. Rudloff, and R. F. Hartl, “Comparison
of anticipatory algorithms for a dial-a-ride problem,” European
Journal of Operational Research, vol. 301, no. 2, pp. 591–608, 2022.
[119] P. Ruiz, M. Seredynski, ´
A. Torn´
e, and B. Dorronsoro, “A digital
twin for bus operation in public urban transportation systems,” in
Big Data Intelligence and Computing, C.-H. Hsu, M. Xu, H. Cao,
H. Baghban, and A. B. M. Shawkat Ali, Eds. Singapore: Springer
Nature Singapore, 2023, pp. 40–52.
[120] S. Hasan, J. Gorospe, S. Girs, A. A. G ´
omez, and E. Uhlemann,
“Platoonsafe: An integrated simulation tool for evaluating platoon
safety,” IEEE Open Journal of Intelligent Transportation Systems,
vol. 4, pp. 325–347, 2023.
[121] J. Khalil, D. Yan, L. Yuan, M. Jafarzadehfadaki, S. Adhikari, V. P.
Sisiopiku, and Z. Jiang, “Realistic urban traffic simulation with
ride-hailing services: A revisit to network kernel density estimation
(systems paper),” in Proceedings of the 30th International Conference
on Advances in Geographic Information Systems, ser. SIGSPATIAL
’22. New York, NY, USA: Association for Computing Machinery,
2022. [Online]. Available: https://doi.org/10.1145/3557915.3560963
[122] X. Xu, B. Shen, S. Ding, G. Srivastava, M. Bilal, M. R. Khosravi,
V. G. Menon, M. A. Jan, and M. Wang, “Service offloading with
deep q-network for digital twinning-empowered internet of vehicles
in edge computing,” IEEE Transactions on Industrial Informatics,
vol. 18, no. 2, pp. 1414–1423, 2022.
[123] X. Xu, Z. Liu, M. Bilal, S. Vimal, and H. Song, “Computation
offloading and service caching for intelligent transportation systems
with digital twin,” IEEE Transactions on Intelligent Transportation
Systems, vol. 23, no. 11, pp. 20 757–20 772, 2022.
[124] K. Zhang, J. Cao, and Y. Zhang, “Adaptive digital twin and mul-
tiagent deep reinforcement learning for vehicular edge computing
and networks,” IEEE Transactions on Industrial Informatics, vol. 18,
no. 2, pp. 1405–1413, 2022.
[125] T. Liu, L. Tang, W. Wang, X. He, Q. Chen, X. Zeng, and H. Jiang,
“Resource allocation in dt-assisted internet of vehicles via edge
intelligent cooperation,” IEEE Internet of Things Journal, vol. 9,
no. 18, pp. 17 608–17 626, 2022.
[126] X. Yuan, J. Chen, N. Zhang, J. Ni, F. R. Yu, and V. C. M. Leung,
“Digital twin-driven vehicular task offloading and irs configuration
in the internet of vehicles,” IEEE Transactions on Intelligent Trans-
portation Systems, pp. 1–15, 2022.
[127] Y. Liu, A. Yang, Q. Zeng, Y. Sun, J. Gao, and Z. Lv, “Task scheduling
of real-time traffic information processing based on digital twins,”
IEEE Transactions on Intelligent Transportation Systems, pp. 1–9,
2022.
[128] Y. Hui, G. Zhao, C. Li, N. Cheng, Z. Yin, T. H. Luan, and X. Xiao,
“Digital twins enabled on-demand matching for multi-task federated
learning in hetvnets,” vol. 72, no. 2, 2023, pp. 2352–2364.
[129] J. Zheng, T. H. Luan, Y. Hui, Z. Yin, N. Cheng, L. Gao, and L. X.
Cai, “Digital twin empowered heterogeneous network selection in
vehicular networks with knowledge transfer,” IEEE Transactions on
Vehicular Technology, vol. 71, no. 11, pp. 12154–12 168, 2022.
[130] F. Dembski, U. W¨
ossner, M. Letzgus, M. Ruddat, and C. Yamu,
“Urban digital twins for smart cities and citizens: The case study of
herrenberg, germany,” Sustainability, vol. 12, no. 6, 2020. [Online].
Available: https://www.mdpi.com/2071-1050/12/6/2307
[131] J. Schaniel, “Towards a digital twin of seychelles’ road transport
system,” Master’s thesis, IVT, ETH Zurich, Example City, CA, March
2022, available at https://www.research-collection.ethz.ch/handle/20.
500.11850/549150.
[132] S. Or, “Transportation modeling: Building future-
proof mobility,” 2022, accessed: 2024-12-12. [On-
VOLUME , 29
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Mart´
ınez et al.: Sustainable Intelligent Transportation Systems via Digital Twins: A Contextualized Survey
line]. Available: https://blog.ptvgroup.com/en/modeling-planning/
ptv-eguide-transportation-modeling/
[133] R. Hakimian, “Tfl to create digital twin of london underground to
monitor track and tunnels,” 2022, accessed: 2024-12-12. [Online].
Available: https://www.newcivilengineer.com/
[134] V. M. G. Mart´
ınez, M. R. N. Ribeiro, and V. F. S. Mota, “Wi-
fi faces the new wireless ecosystem: a critical review,” Annals of
Telecommunications, vol. 79, no. 5, pp. 397–413, Jun 2024.
[135] G. Naik, B. Choudhury, and J.-M. Park, “Ieee 802.11bd & 5g nr
v2x: Evolution of radio access technologies for v2x communications,”
IEEE Access, vol. 7, pp. 70 169–70 184, 2019.
[136] G. A. Association, “Conclusions and recommendations for commu-
nications service providers supporting road operator priorities and
expectations,” Tech. Rep., 2023.
[137] L. Bao, Q. Wang, and Y. Jiang, “Review of digital twin for intelligent
transportation system,” in 2021 International Conference on Infor-
mation Control, Electrical Engineering and Rail Transit (ICEERT),
2021, pp. 309–315.
[138] M. K. Sengul, C. Tarhan, and V. Tecim, “Application of intelligent
transportation system data using big data technologies,” in 2022 Inno-
vations in Intelligent Systems and Applications Conference (ASYU),
2022, pp. 1–6.
[139] C. Schmidt, F. Volz, L. Stojanovic, and G. Sutschet, “Increasing
interoperability between digital twin standards and specifications:
Transformation of dtdl to aas,” Sensors, vol. 23, no. 18, 2023.
[140] D. Lehner, J. Pfeiffer, E.-F. Tinsel, M. M. Strljic, S. Sint, M. Vier-
hauser, A. Wortmann, and M. Wimmer, “Digital twin platforms:
Requirements, capabilities, and future prospects,” IEEE Software,
vol. 39, no. 2, pp. 53–61, 2022.
[141] T. Kreuzer, P. Papapetrou, and J. Zdravkovic, “Artificial intelligence
in digital twins—a systematic literature review,” Data & Knowledge
Engineering, vol. 151, p. 102304, 2024.
[142] M. Veres and M. Moussa, “Deep learning for intelligent transporta-
tion systems: A survey of emerging trends,” IEEE Transactions on
Intelligent Transportation Systems, vol. 21, no. 8, pp. 3152–3168,
2020.
[143] W. Booyse, D. N. Wilke, and S. Heyns, “Deep digital twins for
detection, diagnostics and prognostics,” Mechanical Systems and
Signal Processing, vol. 140, p. 106612, 2020.
[144] A. Sharma, E. Kosasih, J. Zhang, A. Brintrup, and A. Calinescu,
“Digital twins: State of the art theory and practice, challenges,
and open research questions,” Journal of Industrial Information
Integration, vol. 30, p. 100383, 2022.
[145] J. R. Montoya-Torres, S. Moreno, W. J. Guerrero, and G. Mej´
ıa, “Big
data analytics and intelligent transportation systems**this work was
supported by universidad de la sabana, colombia (grant ingphd-10-
2019).” IFAC-PapersOnLine, vol. 54, no. 2, pp. 216–220, 2021, 16th
IFAC Symposium on Control in Transportation Systems CTS 2021.
[146] M. Aloqaily, O. Bouachir, F. Karray, I. Al Ridhawi, and A. E. Saddik,
“Integrating digital twin and advanced intelligent technologies to re-
alize the metaverse,” IEEE Consumer Electronics Magazine, vol. 12,
no. 6, pp. 47–55, 2023.
[147] Y. Yuan and S. Chandrasekaran, “Laying the foundation for
extended reality,” February 2023. [Online]. Available: https:
//spectrum.ieee.org/laying-foundation-for-extended-reality
[148] ITU-T Focus Group on metaverse, “Exploring the metaverse: op-
portunities and challenges,” ITU-T, Tech. Rep. FG-MV D.WG1-01,
2023.
[149] S. Mihai, M. Yaqoob, D. V. Hung, W. Davis, P. Towakel, M. Raza,
M. Karamanoglu, B. Barn, D. Shetve, R. V. Prasad, H. Venkataraman,
R. Trestian, and H. X. Nguyen, “Digital twins: A survey on enabling
technologies, challenges, trends and future prospects,” IEEE Commu-
nications Surveys & Tutorials, vol. 24, no. 4, pp. 2255–2291, 2022.
[150] F. Tang, X. Chen, T. K. Rodrigues, M. Zhao, and N. Kato, “Survey
on digital twin edge networks (diten) toward 6g,” IEEE Open Journal
of the Communications Society, vol. 3, pp. 1360–1381, 2022.
[151] X. Wang, Z. Huang, S. Zheng, R. Yu, and M. Pan, “Unpredictability
of digital twin for connected vehicles,” China Communications,
vol. 20, no. 2, pp. 26–45, 2023.
[152] O. Ercetin, “Computational and communication aspects of digital
twin: An information theoretical perspective,” IEEE Communications
Letters, vol. 27, no. 2, pp. 492–496, 2023.
[153] Y. Song, C. Guo, P. Xu, and J. Wang, “Design of deterministic
transmission framework for distributed power system based on digital
twin,” in 2021 IEEE 5th Conference on Energy Internet and Energy
System Integration (EI2), 2021, pp. 3391–3395.
[154] T. Q. Duong, D. Van Huynh, S. R. Khosravirad, V. Sharma, O. A.
Dobre, and H. Shin, “From digital twin to metaverse: The role
of 6g ultra-reliable and low-latency communications with multi-tier
computing,” IEEE Wireless Communications, vol. 30, no. 3, pp. 140–
146, 2023.
[155] A. Nasrallah, A. S. Thyagaturu, Z. Alharbi, C. Wang, X. Shao,
M. Reisslein, and H. ElBakoury, “Ultra-low latency networks: The
ieee tsn and ietf detnet standards and related 5g ull research,” IEEE
Communications Surveys & Tutorials, vol. 21, no. 1, pp. 88–145,
2019.
[156] N. Finn, J.-Y. L. Boudec, E. Mohammadpour, J. Zhang, and B. Varga,
“Deterministic Networking Bounded Latency,” RFC 9320, nov 2022.
[157] K. Cao, L. Zhang, and C. Zhou, “Design for low latency and high
reliability: Opc-ua based production line for product traceability and
customized service,” in 2021 International Conference on Intelligent
Technology and Embedded Systems (ICITES), 2021, pp. 71–76.
[158] V. M. G. Mart´
ınez, D. A. do Nascimento, D. R. Campelo, R. C.
Mello, and M. R. N. Ribeiro, “On latency bounds for ros messaging
in cloud-enabled mobile robots,” in 2022 IEEE 8th World Forum on
Internet of Things (WF-IoT), 2022, pp. 1–6.
[159] R. Moreira, R. S. Villac¸a, M. R. Ribeiro, J. S. Martins, J. H.
Corrˆ
ea, T. C. Carvalho, and F. de Oliveira Silva, “An intelligent
native network slicing security architecture empowered by federated
learning,” Future Generation Computer Systems, vol. 163, p. 107537,
2025.
[160] Y. Wang, Z. Su, S. Guo, M. Dai, T. H. Luan, and Y. Liu, “A
survey on digital twins: Architecture, enabling technologies, security
and privacy, and future prospects,” IEEE Internet of Things Journal,
vol. 10, no. 17, pp. 14 965–14 987, 2023.
[161] E. Karaarslan and M. Babiker, “Digital twin security threats and
countermeasures: An introduction,” in 2021 International Conference
on Information Security and Cryptology, 2021, pp. 7–11.
[162] C. He, T. H. Luan, R. Lu, Z. Su, and M. Dong, “Security and privacy
in vehicular digital twin networks: Challenges and solutions,” IEEE
Wireless Communications, vol. 30, no. 4, pp. 154–160, 2023.
[163] C. L. Stergiou, E. Bompoli, and K. E. Psannis, “Security and privacy
issues in iot-based big data cloud systems in a digital twin scenario,”
Applied Sciences, vol. 13, no. 2, 2023.
[164] S. Suhail, S. Zeadally, R. Jurdak, R. Hussain, R. Matuleviˇ
cius, and
D. Svetinovic, “Security attacks and solutions for digital twins,” Tech.
Rep., 2022.
[165] R. Jabbar, E. Dhib, A. B. Said, M. Krichen, N. Fetais, E. Zaidan, and
K. Barkaoui, “Blockchain technology for intelligent transportation
systems: A systematic literature review,” IEEE Access, vol. 10, pp.
20 995–21 031, 2022.
[166] H. Feng, D. Chen, and Z. Lv, “Blockchain in digital twins-based
vehicle management in vanets,” IEEE Transactions on Intelligent
Transportation Systems, vol. 23, no. 10, pp. 19613–19 623, 2022.
[167] H. Chai, S. Leng, J. He, K. Zhang, and B. Cheng, “Cyber-
chain: Cybertwin empowered blockchain for lightweight and privacy-
preserving authentication in internet of vehicles,” IEEE Transactions
on Vehicular Technology, vol. 71, no. 5, pp. 4620–4631, 2022.
[168] K. Y. H. Lim, P. Zheng, and C.-H. Chen, “A state-of-the-art survey of
digital twin: techniques, engineering product lifecycle management
and business innovation perspectives,” Journal of Intelligent Manu-
facturing, vol. 31, no. 6, pp. 1313–1337, Aug 2020.
[169] A. Agrawal, R. Thiel, P. Jain, V. Singh, and M. Fischer, “Digital twin:
Where do humans fit in?” Automation in Construction, vol. 148, p.
104749, 2023.
[170] P. G. Larsen, J. Fitzgerald, and J. Woodcock, “How do we engineer
trustworthy digital twins?” Research Directions: Cyber-Physical Sys-
tems, vol. 1, p. e3, 2023.
[171] N. Tripathi, H. Hietala, Y. Xu, and R. Liyanage, “Stakeholders
collaborations, challenges and emerging concepts in digital twin
ecosystems,” Information and Software Technology, vol. 169, p.
107424, 2024.
[172] R. Ehwi, “Engaging stakeholders in digital infrastructure projects,”
2019.
[173] C. Kober, F. G. Medina, M. Benfer, J. P. Wulfsberg, V. Martinez, and
G. Lanza, “Digital twin stakeholder communication: Characteristics,
challenges, and best practices,” Computers in Industry, vol. 161, p.
104135, 2024.
30 VOLUME ,
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
[174] M. Faherty, G. McArdle, and P. Tafidis, “Envisioning sustainable ur-
ban mobility in dublin: Stakeholder perspectives and system dynam-
ics,” Transportation Research Interdisciplinary Perspectives, vol. 28,
p. 101280, 2024.
[175] J. Zhang, S. Li, and Y. Wang, “Shaping a smart transportation system
for sustainable value co-creation,” Information Systems Frontiers,
vol. 25, no. 1, pp. 365–380, Feb 2023.
[176] E. Marcucci, V. Gatta, M. Le Pira, L. Hansson, and S. Br˚
athen,
“Digital twins: A critical discussion on their potential for support-
ing policy-making and planning in urban logistics,” Sustainability,
vol. 12, no. 24, 2020.
VICTOR M. G. MARTINEZ received his B.Sc.
degree in telecommunications engineering from
the Instituto Superior Polit´
ecnico Jos´
e Antonio
Echeverr´
ıa (CUJAE, Cuba) in 2011, and his M.Sc.
degree in electrical engineering from the Federal
University of Esp´
ırito Santo (UFES, Brazil) in
2019. He is currently working towards a Ph.D.
degree in electrical engineering at UFES. His re-
search interests include software-defined networks,
network performance analysis, and cyber-physical
systems.
DIVANILSON R. CAMPELO (Member, IEEE)
received the degree in electrical engineering from
the Universidade Federal de Pernambuco (UFPE),
Recife, Brazil, in 1998, and the M.Sc. and Doc-
toral degrees in electrical engineering from the
Universidade Estadual de Campinas (UNICAMP),
Campinas, Brazil, in 2001 and 2006, respectively.
He was a Visiting Assistant Professor with the
Electrical Engineering Department, at Stanford
University, Stanford, CA, USA, from September
2008 to December 2009. He is an Associate Pro-
fessor of computer science and computer engineering with the Centro de
Inform´
atica (CIn), UFPE. His research interests include automotive and
vehicular networking, connected objects, broadband access networks, and
cybersecurity.
MOISES R. N. RIBEIRO received the B.Sc. de-
gree in electrical engineering from the Instituto
Nacional de Telecomunicac¸ ˜
oes, Brazil, in 1992,
the M.Sc. degree in telecommunications from the
Universidade Estadual de Campinas, Brazil, in
1996, and the Ph.D. degree from the University
of Essex, U.K., in 2002. In 1995, he joined
the Department of Electrical Engineering, Federal
University of Esp´
ırito Santo. He was a Visiting
Professor with the Photonics and Networking Re-
search Laboratory, at Stanford University from
2010 to 2011. His research interests include fiber optic communication,
sensor devices, systems, and networks.
VOLUME , 31
This article has been accepted for publication in IEEE Open Journal of Intelligent Transportation Systems. This is the author's version which has not been fully edited and
content may change prior to final publication. Citation information: DOI 10.1109/OJITS.2025.3553696
This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
