Artificial Intelligence Revolutionising the Automotive Sector: A

Comprehensive Review of Current Insights, Challenges, and Future Scope

AUTHOR(S)

MN Hossain, Md Abdur Rahim, MM Rahman, D Ramasamy

PUBLICATION DATE

01-01-2025

HANDLE

10779/DRO/DU:28852088.v1

Downloaded from Deakin University’s Figshare repository

Deakin University CRICOS Provider Code: 00113B

 Tech Science Press

Doi:10.32604/cmc.2025.061749

REVIEW

Artificial Intelligence Revolutionising the Automotive Sector: A Comprehensive

Review of Current Insights, Challenges, and Future Scope
Md Naeem Hossain1 , Md. Abdur Rahim2 , Md Mustafizur Rahman1,3, * and Devarajan Ramasamy1
1
Faculty of Mechanical and Automotive Engineering Technology, Universiti Malaysia Pahang Al-Sultan Abdullah, Pekan, 26600,
Malaysia
2
Institute for Intelligent Systems Research and Innovation (ISSRI), Deakin University, Warun Ponds, VIC, 3216, Australia
3
Automotive Engineering Centre, Universiti Malaysia Pahang Al-Sultan Abdullah, Pekan, 26600, Malaysia
*Corresponding Author: Md Mustafizur Rahman. Email: [email protected]
Received: 02 December 2024; Accepted: 06 February 2025; Published: 06 March 2025

ABSTRACT: The automotive sector is crucial in modern society, facilitating essential transportation needs across
personal, commercial, and logistical domains while significantly contributing to national economic development and
employment generation. The transformative impact of Artificial Intelligence (AI) has revolutionised multiple facets of
the automotive industry, encompassing intelligent manufacturing processes, diagnostic systems, control mechanisms,
supply chain operations, customer service platforms, and traffic management solutions. While extensive research
exists on the above aspects of AI applications in automotive contexts, there is a compelling need to synthesise this
knowledge comprehensively to guide and inspire future research. This review introduces a novel taxonomic framework
that provides a holistic perspective on AI integration into the automotive sector, focusing on next-generation AI
methods and their critical implementation aspects. Additionally, the proposed conceptual framework for real-time
condition monitoring of electric vehicle subsystems delivers actionable maintenance recommendations to stakeholders,
addressing a critical gap in the field. The review highlights that AI has significantly expedited the development
of autonomous vehicles regarding navigation, decision-making, and safety features through the use of advanced
algorithms and deep learning structures. Furthermore, it identifies advanced driver assistance systems, vehicle health
monitoring, and predictive maintenance as the most impactful AI applications, transforming operational safety and
maintenance efficiency in modern automotive technologies. The work is beneficial to understanding the various use
cases of AI in the different automotive domains, where AI maintains a state-of-the-art for sector-specific applications,
providing a strong foundation for meeting Industry 4.0 needs and encouraging AI use among more nascent industry
segments. The current work is intended to consolidate previous works while shedding some light on future research
directions in promoting further growth of AI-based innovations in the scope of automotive applications.

KEYWORDS: Artificial intelligence; AI techniques; automotive sector; autonomous vehicle; decision-making; VHMS

1 Introduction
In the fourth industrial revolution era, artificial intelligence (AI) significantly impacts the world in
resolving hitherto intractable issues from different sectors. AI mimics human behaviour like planning, rea-
soning, learning, problem-solving, thinking, perception, and decision-making in machines. This cognitive
capability is achieved through human intelligence simulation utilising data with feasible algorithms [1,2].
As a result, AI can simplify complex tasks to minimise errors more efficiently, whereas humans are more
prone to errors and require more time to complete the same tasks. Therefore, AI adoption in different sectors

Copyright © 2025 The Authors. Published by Tech Science Press.

This work is licensed under a Creative Commons Attribution 4.0 International License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
3644 Comput Mater Contin. 2025;82(3)

is increasing dramatically [3], i.e., the automotive sector [4], healthcare sector [5], power sector [6], real
estate sector [7], entertainment and gaming sector [8], fast-moving consumer goods sector [9], education
sector [10], e-commerce and retail shop sector [11], agriculture sector [12], cybersecurity [13], finance and
banking sector [14], and so forth. Since the number of AI-adopted sectors is rising with time, its market is
also expanding. VTT Technical Research Centre discusses these matters in their technical paper. They found
that the AI market price in 2019 was $39.9 billion, and market analysts predict that this will increase to $733.7
billion in 2027 [15].
Moreover, based on this evidence, it is clear that AI’s significance in different sectors is rising due
to its massive beneficial prospects and the outstanding growth of related research activities. Similarly,
its importance is also visible in the automotive industry in minimising issues that arise and maximising
outcomes intelligently. For example, intelligent fault diagnosis systems [16], intelligent manufacturing
processes [17], intelligent control systems [18], intelligent supply chains, sales and service systems [19],
intelligent traffic management systems [20], etc. Adopting AI for the abovementioned purposes makes the
automotive sector more intelligent and sophisticated. Curiosity about AI is expanding in the automotive
domain because it holds the key to the new future. It can extract essential information without the direct
intervention of humans and utilise these to run any operation/activity smoothly to gain desired outcomes
efficiently besides minimum losses in the automotive sector. As a result, the AI deployment rate in this
sector for various purposes is also praiseworthy [21]. According to the report of Precedence Research,
the worldwide automotive AI market generated revenue of $3.22 billion in 2023 and is expected to grow
from $3.87 billion in 2024 to $35.71 billion by 2033. This remarkable expansion represents a compound
annual growth rate of 28% over the nine-year forecast period [22]. Enhancing it in this sector will accelerate
economic growth worldwide and improve desired functionalities. However, nuTonomy, AutoX, Optimus
Ride, Drive.ai, Waymo, Zoox, CarVi, Nauto, Rethink Robotics, DataRPM, and many other organisations
have been using AI technology in the last few years to shape the business context of the automotive sector
and move it ahead [3].
Furthermore, the AI technology in this sector acts as a machine-enabled intelligent system that can
perceive, learn, analyse, and comprehend the outcomes by following human cognition activity. To ensure
the facility of this technology, it usually comprises algorithms with an arrangement of instructions loaded
in a machine-like computer, which transforms input data into meaningful output [23]. In this instance,
traditionally, it is not confined to machine learning (ML) algorithms used in the automotive sector as AI
to expedite the desired work without complexity [24]. Artificial neural networks (ANN), deep learning
(DL), reinforcement learning (RL), fuzzy logic, and so on are also currently used to mimic human cognition
for better results with utmost efficiency [23]. In previous times, due to AI implementation for different
purposes in this sector, various challenging issues arose during the operational period, hindering continuity
and reducing overall efficiency. In this case, data quantity context (big data), large-scale implementation,
real-time results, processing speed, systems downtimes, costing, and many other issues usually appear
challenging. However, the respected researchers continue their exceptional work to mitigate these challenges
by perfectly performing the AI system [23,24].
The recent progress of AI adoption and its potential prospects for the future in the automotive sector
motivated us to gather valuable information to assist future research work and step ahead of this sector with
more sophisticated AI technology. Numerous types of exceptional research have already been carried out
independently by focusing on particular issues. For example, Fedullo et al. [25] reviewed the application of
AI techniques in the automotive sector, with a focus on innovative measurement systems, advanced driver
assistance systems (ADAS), Internet of Things (IoT), and intelligent industrial systems to improve road safety,
predictive maintenance (PdM), and build the intelligent automotive factory of the future. Subsequently,
Comput Mater Contin. 2025;82(3) 3645

to enhance traffic control and vehicle communication in connected vehicle frameworks, Rana et al. [26]
summarised a review showing a broad range of AI applications in the automotive industry beyond just ADAS,
including car emissions, PdM, and security. However, the researchers suggested future studies on network
architecture, connectivity, and performance metrics related to automotive AI and ML applications. Addi-
tionally, Ammal and coworkers [27] provided an overview of AI and sensor technology to develop innovative
products and applications in the automotive industry, reducing human errors such as aggressive driving,
accidents, and traffic collisions. Moreover, a literature study by Vermesan and colleagues [28] summarised
that AI technologies enable intelligent functions and optimisation for electric-connected autonomous
and shared vehicles to support sustainable green mobility. Nevertheless, decision-making mechanisms in
autonomous vehicles need to be developed using AI, machine learning, deep learning, and other advanced
techniques to make the processes more reliable and resilient. Despite recent review works offering valuable
insights into the use of AI methods in specific aspects of the automotive sector, such as control systems,
vehicle diagnostics, traffic management, autonomous driving, and accident prevention, these studies often
approach applications in isolation. This fragmented analysis fails to provide a comprehensive understanding
of how AI techniques interact, complement, or conflict with one another when integrated into a cohesive
automotive ecosystem. Additionally, cross-functional challenges such as interoperability issues, data sharing,
and real-time decision-making within interconnected automotive domains remain underexplored, limiting
the potential for a unified and optimised implementation of AI in the industry. To address this research
gap, it is essential to develop a holistic perspective that integrates various AI techniques, such as machine
learning, deep learning, computer vision, and natural language processing, across all application areas
within the automotive sector. This integrated approach would enable a comprehensive understanding of
the combined potential of these technologies, uncover synergistic opportunities, and provide clarity on
overarching challenges. By fostering such a consolidated framework, the automotive industry can move
toward building truly intelligent, resilient, and optimised systems. Therefore, this review aims to bridge
this gap by holistically examining the state-of-the-art AI techniques, assessing their collective impact, and
outlining a roadmap for future research to address these integrative challenges in the automotive industry.
In particular, this article infers:
i. Fundamental details about AI technology for the automotive sector;
ii. As pictorial maps, a proposed taxonomy of AI in the automotive field;
iii. Dig out the inclusive information of different AI approaches and their applications in the automotive
sector;
iv. Profound analysis of benefits, associated challenges, and future application scope of AI in this sector.
Ultimately, we anticipate that this article’s outcome will represent the credible capture of AI technology’s
diversified utilisation and advancement in the automotive sector. Additionally, it will assist as a robust
foundation for fulfilling the demand of Industry 4.0 requirements and intensifying the adoption of AI
technology in the underprivileged branches of this sector.
The remaining portions of this paper are presented in the following order. Firstly, Section 2 presents
a history of AI adoption in the automotive field and utilises this information to portray the taxonomy
in Section 3, besides the extensive discussion. Then, Section 4 describes the automotive-based AI conceptual
framework, and Section 5 outlines the challenges and potential areas for future AI research in the automotive
field. Finally, Section 6 ends this review by emphasising significant contributions and observations.
3646 Comput Mater Contin. 2025;82(3)

2 Historical Background
Due to the pressing need for the quick processing of large amounts of data, AI’s importance in daily life
is becoming indispensable. However, this stage of AI is not achieved in a day or a year. It evolved over 70
years and started from the 1940s; indeed, 1942 through Runaround was a short story written by Isaac Asimov,
the American Science Fiction writer, where he plotted a robot that allowed it to work for human beings based
on some fundamental rules. Later, it inspired scientists and was first coined as “Artificial Intelligence” in 1956
by John McCarthy and Marvin Minsky through many transformations and changes [29]. In 1961, General
Motors installed an AI-based industrial robot to advance the automotive industry for managing die castings
smoothly. Later, they inspired and enhanced the Lordstown assembly plant using this technology in 1969.
They established an AI-based automated vehicular assembly plant in the mid-1970s with MIT to expedite the
vehicle manufacturing process [30]. After the industrial cases, the automotive sector also enhanced drive-in
highways using this technology.
In 1979, Tsugawa and his research team [31] proposed driverless vehicles as autonomous vehicles (AVs)
by recognising road patterns using TV cameras. Such autonomous vehicles could move within 30 Km/hr
without driver assistance in different road environments. Also, in the early 1980s, autonomous land vehicles
with AI-contained high-performance computing facilities were introduced by DARPA to run the vehicle
over the highway at a maximum 45 mph speed [32]. The AI system was confined to the purposes mentioned
above for the automotive sector, and it was extended to diagnosing vehicles like an expert system by 1986 [33].
Recently, this concept has attained great attention and is being applied in various cases to diagnose faults
in the automotive sector to minimise downtime and increase operational time, which is discussed later
elaborately. However, in the early 1990s, researchers began to realise the significance of intelligent traffic
management and the employment of AI technology in this sector, which is eloquently discussed in [34].
Besides these, this technology also showed praiseworthy outcomes for intellectual vehicular performance
controlling systems in 1995. Furthermore, this work inspired and upgraded AI technology to achieve
supremacy for regulating vehicular performance more sophisticatedly [35].
It is noticeable that from the beginning, AI technology’s footprint significantly appeared in the automo-
tive field for manufacturing, autonomous driving, traffic management, vehicle performance regulating, and
fault diagnosing purposes [23,36]. Before 2000, AI’s evolution in this sector was in its initial stages compared
to other industries. The giant leap happened in 2009, as over the last decade, Google, Tesla, General Motors,
Volvo, Intel Corporation, IBM Corporation, and so forth, renowned organisations have undertaken many
projects to employ AI technology with a more sophisticated approach for advancing the automotive sector
to a great extent [37–39]. However, the adoption of AI is increasing outstandingly, and it is necessary to
portray the taxonomy as a map of detailed implementation in the automotive field, described extensively in
the upcoming section.

3 Taxonomy of AI in the Automotive Sector

AI applications in the automotive industry have made our lives easier and more dynamic. Therefore,
it is unavoidable that all the related aspects of AI in the automotive sector be depicted in an integrative
format. For this reason, a taxonomy has been developed (Fig. 1) to give a clear view of AI’s employment in
the automotive sector.

3.1 Classification of AI Systems

Through human-inspired, analytical, and humanised, depending on the intelligence exhibition (emo-
tional, cognitive, and social) and the evolutionary stage of AI in the automotive sector, it is divided into two
distinct groups: Type I and Type II [29,40].
Comput Mater Contin. 2025;82(3) 3647

Type I

Reactive Limited Theory of

Self-Aware
Machines Theory Mind
Classification
of AI System Type II

ANI AGI ASI

SVM, KNN, DT, RF K-Means Clustering, Hierarchical

Supervised Unsupervised Semi-Supervised Reinforcement

CNN, RNN, GRU SI GA ACO

ANN ML DL Heuristic Metaheuristic

Technical
AI in Statistical Game Theoretic Fuzzy Hybrid Intelligent Computer Vision
Intelligent Systems
Automotive

NLP Gesture Speech

Service Computing Network Security

Intelligence Intelligence Intelligence Intelligence

Intelligent
Applications ADAS VHMS Manufacturing Supply Chain PdM

Adaptive Cruise Control Collision Avoidance Automatic Parking

Figure 1: Proposed taxonomy of AI in the automotive sector

3.1.1 Type I
Under this category, the employed AI technology in the automotive sector could behave like a human
mind exhibiting the ability to think and feel based on the functionality of the AI system. Moreover, it is
classified into four subcategories [41] and described as follows:
Reactive Machines
The AI system does not depend on memory or utilise previous experiences and decides only based on
the current situation or existing scenario [42]. As a result, all control and decisions in the system might
be conducted by light processing using current data. So, in this case, low-priced and limited resources
are required to get better output, but in some cases may be trapped in complex environments [43]. For
example, Mr Roberts and his research team proposed [44] an autonomous underground mining vehicle
to ensure operator safety, reducing operational complexity using reactive machines AI. The robust reactive
wall-following behaviour was the control architecture for 30 tons Load-Haul-Dump trucks.
3648 Comput Mater Contin. 2025;82(3)

Limited-Memory Machines
The AI system can acquire knowledge and learn from historical data as past experiences, stored data,
actions, and learnings to make subsequent decisions. That means this combination of observational data and
preprogrammed knowledge is used to make the necessary decision. In this case, the past data is not stored
for an extended period [40–42]. AVs are a real-world example of limited-memory machines. It can read
their environment to discover patterns and changes in external elements to adapt and learn them, besides
observing and understanding how to drive human-operated vehicles efficiently. Also, it can detect other
vehicles and pedestrians in their line of sight [45]. Previously, such feats may take up to 100 s, but this time
has drastically lowered due to recent technology and software advancements, i.e., deep learning and machine
learning [40].

Theory of Mind
The theory of mind AI is a more advanced stage of AI systems than the previous two systems, which
can be described as a human mind. As a result, this system may study human emotional complexities
besides behavioural patterns and make predictions about their emotions, intentions, desires, thoughts, and
beliefs [41]. In this case, the significant aspect involves grouping entities based on similarity, proximity,
symmetry, continuity, common fate, etc. [46]. Thus, drones and AVs might benefit from this AI system
because it could avoid physical harm to ordinary people and pedestrians by predicting the movements as
humanity’s collective ability [47]. Also, in automotive manufacturing environments, it could help measure
employees’ perceptions and behaviour to expedite the manufacturing process [48].

Self-Aware AI
The last step of AI system development is self-aware AI, which currently exhibits hypothetically, and
in this case, will attain human-level perception, emotions, desires, and independent decision-making ability
based on its own moral compass and philosophical approach. Respected researchers are working on adopting
it in the automotive sector, such as AVs, autonomous manufacturing systems, and industrial management;
it has a broad scope to continue safe and effective operations in all cases [49].

3.1.2 Type II
In this context, AI technology classified based on capabilities and technology in the automotive sector
could be organised into three subcategories [49], which are discussed as follows:
Narrow Intelligence
Narrow intelligence can be defined as not exhibiting intelligent functionality beyond the particular
application domain precisely on what they have been programmed to do [41]. Widely, it performs efficiently
on specific tasks based on the given initial training, but generalisation ability is missing except for these
particular tasks [48]. Such limited focus capabilities represent AI’s narrowness, also termed “weak AI” [41].
However, the examples of narrow intelligence are related to afore-discussed reactive machines and limited-
memory machines AI. Also, in some cases, the theory of the mind machine’s rudimentary examples is
involved [40].

General Intelligence
Artificial general intelligence can be defined as a system that not only performs like a weak AI. However,
that system also has the capabilities or functionalities to understand, learn, perceive, and make decisions
Comput Mater Contin. 2025;82(3) 3649

entirely the same as human beings [46]. That means the system in the automotive sector could perform any
operation like a human without human fatigue besides error [40]. Because of this, it is also called “strong
AI.” Moreover, though it does not exist, the continuous development and advancement of narrow AI in
the automotive sector minimised the distance between humans and AI machines’ capabilities. Therefore,
hopefully, it will lead to general intelligence very soon [41].

Super Intelligence
It is still in the conceptual stages where the AI system in the automotive sector could have higher
cognitive functional capabilities than the human brain. However, since it exists hypothetically, it is hoped
that the system may have the total ability to perform moving and transportation services due to enhanced
memory besides significantly faster data processing, handling, and decision-making abilities [23]. Table 1
presents the fundamental differences between the different categories of AI in various aspects.

Table 1: Comparison between AI classes

Reference AI class Definition Capabilities Learning ability Challenges

Hassani Reactive Simple AI that Basic No memory, cannot Limited to specific
et al. [42] machines responds to current decision-making in improve or adapt tasks; lacks
inputs without real-time based on over time adaptability and
memory pre-defined rules learning
Shi et al. [50] Limited- AI that can use past Can learn from Can learn from past Limited by data
memory experiences to historical data and data but not in processing power,
machines make decisions adjust behaviour real-time; retains memory fades over
accordingly short-term memory time
Nori et al. [51] Theory of AI is capable of Advanced Hypothetical: Currently
mind understanding decision-making would require speculative, it
emotions, beliefs, with interaction adaptive learning of requires significant
and intentions based on drivers’ human-like traits advances in
intentions emotional and
situational
understanding
Shao et al. [52] Self-aware AI AI that possesses a Could make Theoretical Largely theoretical;
sense of self, independent self-learning, poses ethical and
potentially decisions, anticipate evolving beyond the existential risks
conscious of its changes, and exhibit initial
existence self-preservation programming
behaviours
Gryth Narrow AI focused on a Highly optimised Pre-trained to Task-specific
et al. [53] intelligence specific, narrow for single-domain particular tasks but limitations: unable
task tasks, limited cannot generalise to handle
generalisation beyond them unfamiliar
ability scenarios effectively
Li et al. [54] General AI that can Flexible Real-time learning Significant
intelligence understand, learn, decision-making, and adaptation technological and
and apply multi-domain could process ethical hurdles
intelligence across problem-solving, multi-modal data limited by current
various domains human-level for better responses AI capabilities
reasoning
(Continued)
3650 Comput Mater Contin. 2025;82(3)

Table 1 (continued)

Reference AI class Definition Capabilities Learning ability Challenges

Chai et al. [55] Super AI that surpasses Autonomous, with Theoretical Highly theoretical;
intelligence human intelligence enhanced continuous significant ethical,
in all aspects problem-solving improvement, security, and safety
and strategic self-modification risks
thinking capabilities

3.2 AI Techniques Used in the Automotive Sector

Commonly used techniques or advanced methods of AI in the automotive field have been applied
to perform human-like or superhuman behaviour to operate or maintain the desired system smoothly.
For manufacturing, autonomous driving, traffic management, vehicle performance regulating, and fault-
diagnosing purposes, various AI approaches and fundamental techniques are currently used to mimic the
human brain’s working principle in the automotive sector, described as follows:

3.2.1 Machine Learning

Machine learning is an algorithm that learns from the data without depending on rules-based pro-
gramming. Currently, the ML technique is one of the most effective AI tools for resolving various issues in
the automotive industry. By exploring the algorithm’s construction and learning mechanisms, ML can study
data patterns to predict the most probable outcomes with high accuracy [56]. Additionally, it shares many
similarities and overlaps with computational statistics. Based on the learning nature of the automotive field,
it is classified into four categories [57], as depicted in Fig. 2; these are explained as follows.

Supervised Learning
The supervised learning approach is an AI algorithm generally trained by the labelled input data to
predict or provide a firm decision. This approach is primarily suitable for classification in the automotive
field besides solving regression problems, and various algorithms are available to achieve the desired
outcomes [57]. In this case, supervised learning includes techniques like support vector machines (SVM),
K-nearest neighbour (KNN), decision trees (DT), random forest (RF), logistic regression, linear regression,
and others. However, in the last decade, the supervised technique has been utilised in the automotive
sector for various purposes, where the system is initially trained by feeding required information and later
run using the testing data. For instance, Harold et al. [58] proposed a powertrain control framework for
hybrid electric vehicles (EV) that uses the supervised learning technique to improve fuel economy, and they
achieved a satisfactory outcome over dynamic environments. Furthermore, due to better performance in
decision-making and classification, this technique is also implemented in the automotive industry to control
manufactured vehicle quality by scrutinising the vehicular body’s panel surfaces [59]. Similarly, last few years,
the demand for supervised learning techniques in the automotive sector has risen significantly for various
purposes, i.e., estimation of moving vehicle’s road friction [60], improving quality of service for intelligent
transport systems [61], performing roundabout manoeuvres for AVs [62], fault detection of unmanned aerial
vehicle (UAV) [63], predictive maintenance and risk management [64,65], supply chain management [66],
and so forth. However, supervised learning is highly dependent on large, high-quality labelled datasets,
which are costly and time-consuming to obtain, particularly for the extensive edge cases seen in real-world
driving environments. This reliance on labelled data creates a contradiction: while supervised learning
Comput Mater Contin. 2025;82(3) 3651

delivers accurate results within a controlled scope, it lacks flexibility, and its performance in dynamic, real-
time situations (e.g., autonomous driving) often necessitates additional support from semi-supervised or
unsupervised learning for comprehensive adaptability.

Supervised Learning

Advanced Driver Assistance Systems

Automated Driving Systems

Predictive Maintenance & Analytics

Personalized Vehicle Functions

Vehicle Detection & Classification

Reinforcement Learning Unsupervised Learning

Autonomous Driving Anomaly Detection in Sensor Data

Vehicle Control & Stability Path Optimization for AVs

Advanced Energy Management Machine Risk Prediction & Decision-Making

Learning
Simulation to Reality Transfer Driver Behavior Analysis
Lane-Keeping Assist Systems Safety-Critical Systems

Semi-Supervised Learning

Radar-Based Object Classification

Human-Car Interaction

Automotive Ethernet Security

Manufacturing Fault Detection

Vehicle Trajectory Prediction

Figure 2: Machine learning techniques for automotive applications

Unsupervised Learning
In machine learning, the unsupervised learning approaches can be defined as the AI algorithm models
or systems that are not supervised based on the labelled training dataset to draw affirmative inferences.
This method is crucial in the automotive industry to uncover hidden data for specific purposes, and
3652 Comput Mater Contin. 2025;82(3)

it currently makes use of a variety of well-liked algorithms, including hierarchical clustering, K-means,
self-organising maps, and many others [67,68]. Moreover, this technique significantly improves in some
cases compared to supervised learning techniques. For example, an unsupervised learning approach in the
automobile manufacturing industry provided praiseworthy performances for detecting the fault and failure
prediction of power transmission systems, pneumatic actuators, and valves [69]. Similarly, the K-means
clustering algorithm was used by Shaeiri and coworkers [70] as an unsupervised learning approach for
vehicular self-maintenance purposes and profiling driver behaviours. They found that the proposed system
successfully assists the user. Besides these purposes, day by day, the implementation of this technique in
the automotive field is increasing considerably, i.e., real-time automobile insurance fraud detection [70],
CAN-based modern vehicle transmission systems [71], enhanced driver assistance systems [72], and so forth.
However, unsupervised learning presents notable disadvantages, particularly the difficulty of validating
results and the risk of identifying irrelevant patterns, as there is no labelled data for model guidance. This
lack of interpretability can hinder safety-critical applications, where understanding the model’s rationale
is essential. Contradictions arise in scenarios like autonomous driving, where accuracy and safety are
paramount. However, unsupervised learning provides flexibility and speed; it often requires subsequent
validation with supervised methods, raising concerns about reliability and the need for hybrid approaches
to balance accuracy with adaptability.

Semi-Supervised Learning
It can be defined as the ML approach, where the system is trained by feeding a small amount of
labelled data with a large quantity of unlabelled data as a mixture of data. In general, it helps to overcome
some difficulties in supervised and unsupervised learning cases where managing a substantial amount
of training data is difficult [73]. Semi-supervised learning outperformed supervised models in vehicle
trajectory prediction by leveraging large amounts of unlabelled data, thus scaling up the training process and
improving accuracy [74]. Additionally, these methods have shown high effectiveness in detecting various
network attacks in automotive Ethernet, achieving impressive detection rates [75]. For example, Hoang and
the research team [76] employed semi-supervised learning to detect various in-vehicle intrusion attacks,
including known attacks like denial of service, fuzzy, and spoofing, as well as unknown attacks. Nevertheless,
one major drawback is the difficulty in handling datasets with few labelled samples, as traditional broad
learning systems struggle to effectively utilise the information between labelled and unlabelled data [77].
Furthermore, the increased interconnectivity in automotive Ethernet introduces new vulnerabilities. While
semi-supervised learning can enhance detection, it is still susceptible to unknown attacks due to the
developing nature of this research area [75].

Reinforcement Learning
Reinforcement learning is a promising ML technique in which the intelligent agent learns from experi-
ences rather than a training dataset and is concerned with initialising the action in the potentially complex
and uncertain environment to maximise reward or outcome by taking a sequence of suitable decisions.
Particularly, in the RL paradigm of the automotive field, an autonomous agent interacts with its environment
to perceive and learn how to improve its performance at a given activity through a series of trials [78].
In addition, its employment proliferation recently appeared and grew dramatically in this field to mitigate
complex issues smartly. Moreover, Navarro et al. [79] found that for developing a simulated environment
and control strategies of an AV, this approach is cost-effective and requires the lowest data solution than the
traditional algorithm. Again, it has extreme significance for vehicle routing problem solutions, and Nazari’s
research team revealed it, proposing an end-to-end framework for self-driven learning [80]. Recently, to get
Comput Mater Contin. 2025;82(3) 3653

better outcomes in this sector, the RL approach enhanced deep reinforcement learning (DRL) by combining
its architecture with ANNs, where software-defined agents enable learning best. As a result, it has performed
praiseworthy in autonomous driving [78,81], UAV navigation [82,83], unmanned surface vehicles [84], and
many others for environment recognition, decision-making, and controlling system advancement. However,
the lack of standard tooling solutions for RL-based function development in the automotive industry remains
challenging [85]. Furthermore, while RL can theoretically enhance driving policies, practical implementation
requires reliable fallback mechanisms and high-fidelity simulators to ensure safety and effective policy
transfer from simulations to real-world applications [86,87].

3.2.2 Deep Learning

The deep learning technique in AI consists of ML algorithms that use multilayered ANNs to mimic
human intelligence by learning and making the best decision, with the most layers acting as the hidden
layer. Its effectiveness has been demonstrated outstandingly for any purpose in the automotive branch,
and implementation is rising dramatically [88]. The significance of DL was realised by Chen’s research
team [89] from Cardiff University, UK, for predictive vehicle maintenance based on maintenance history
and geographical information system (GIS) data. It outperformed based on real-world information than
other traditional approaches and anticipated that it would reduce the fleet management organisations’
maintenance-related burdens. Similarly, Naqvi et al. [90] achieved greater accuracy in mitigating accidental
issues due to inconsistent driver behaviours. The results were much better than those obtained in previous
research using a convolutional neural network (CNN) and DL technique with pupil centre corneal reflection-
based method.
Table 2: Comparison of different deep learning algorithms

Reference DL algorithm Features Application

Ulrich Convolutional Efficient for image-based tasks like Object detection
et al. [91] neural networks recognising pedestrians, vehicles, and and classification
road signs
Paparusso Recurrent neural Models temporal dependencies; useful Trajectory
et al. [92] networks for anticipating driver actions based prediction and
on historical data driver behaviour
Tammewar Deep Q-network Reinforcement learning-based, Autonomous
et al. [93] efficient in path planning and driving control
lane-keeping
Talpaert Deep Learns from trial and error in Autonomous
et al. [94] reinforcement simulated environments, adapting to navigation
learning new driving scenarios
Ren Generative Creates realistic training data, essential Synthetic data
et al. [95] adversarial for rare events in autonomous driving generation for
networks training
Singh Long short-term Efficient in processing time-series Predictive
et al. [96] memory data, detecting anomalies in-vehicle maintenance and
components sensor data
(Continued)
3654 Comput Mater Contin. 2025;82(3)

Table 2 (continued)

Reference DL algorithm Features Application

Wang Autoencoders Used for anomaly detection and Fault detection in
et al. [97] quality control, reducing manual manufacturing
inspection

Again, for demand forecasting and inventory management of automobile spare parts, the DL approach
(modified Adam optimiser with recurrent neural network (RNN)/long short-term memory (LSTM)) per-
formed well and provided less error than available other methods [98]. Moreover, its footprints are also
extensively sketched in remote sensing of UAVs for various purposes in urban, agricultural, and envi-
ronmental contexts. Gated recurrent unit (GRU), Bi-GRU, Bi-LSTM, and DL approaches are prominently
used for fault diagnosis and identification in this sector [99]. Table 2 presents how various DL algorithms
enhance automotive applications like driver assistance systems and traffic management, improving safety and
efficiency. However, these techniques demand significant computational resources and extensive training
data to ensure reliability and safety in real-world scenarios. Therefore, contradictions arise in balancing
the rapid advancements in DL with the automotive industry’s stringent safety and validation requirements,
highlighting the need for ongoing research and development to address these challenges.

3.2.3 Heuristic Techniques

The heuristic technique is an approach for the specific problem that implements the practical method
or numerous shortcut approaches and often gives satisfactory solutions more quickly than other classic
methods [100]. In the automotive sector, day by day, its application is increasing satisfactorily due to its
quick problem-solving features. For instance, decomposition strategies in vehicle routing heuristics can
significantly enhance performance by breaking down significant problems into manageable subproblems,
as demonstrated by the superior performance of route-based decomposition methods over path-based
methods [101]. In addition, the heuristic method with improved multi-choice goal programming is used in
the vehicle-related logistics system to design a green supply chain management system by adopting carbon
regulation as greener practices [102]. Similarly, Pérez’s research team [103] found that the automotive industry
could benefit from the logistic system by minimising transportation costs through minimum touring dis-
tance, loading-unloading operations, and proper decision-making using the heuristic algorithm. However,
the primary disadvantage of heuristic methods is that they often rely on rule-of-thumb approaches that may
not guarantee optimal solutions and can lack robustness across varying conditions, potentially leading to
inconsistent performance. Furthermore, though they provide practical solutions for quick decision-making,
their reliance on approximations and empirical rules may fall short in applications requiring high precision
and predictability.

3.2.4 Metaheuristic Techniques

Metaheuristic approaches emerged as versatile problem-solving tools that operate independently of
specific problems, employing innovative strategies based on heuristic principles to resolve various challenges.
Here, the prefix meta indicates the high-level methodological nature [104]. These techniques, such as those
developed for vehicle routing problems, provide robust solutions to complex combinatorial optimisation
issues by iteratively exploring and exploiting the search space, often outperforming traditional performance
and conceptual originality [105]. Nature-inspired meta-heuristics, in particular, have been successfully
Comput Mater Contin. 2025;82(3) 3655

applied across various domains, including automotive logistics, due to their flexibility and ability to handle
diverse problem sets [106]. Table 3 summarises some innovative approaches to these techniques used in the
automotive sector to optimise a broad range of complex issues perfectly.
Table 3: Comparison between metaheuristic techniques

Reference Aspect Swarm intelligence Genetic algorithm Ant colony

optimisation
[107–109] Definition Mimics the collective Inspired by the A bio-inspired
behaviour of process of natural algorithm that
decentralised systems, selection, involving simulates the
such as bird flocking operations like foraging behaviour
or fish schooling selection, crossover, of ants to find
and mutation optimal paths
[107–110] Key Particle swarm Standard genetic Ant colony system,
algorithms optimisation, artificial algorithm, improved Max-min ant
bee colony, bacterial genetic algorithm with system
foraging optimisation adaptive adjustments
[108,109,111,112] Application Used for vehicle Applied in intelligent Utilised for solving
routing, path vehicle path planning vehicle routing
planning, and to reduce energy problems,
optimisation of consumption and optimising path
logistics improve route planning, and
efficiency enhancing
convergence speed
[108,109,111] Strengths Fast convergence Reduces the number The high success
speed, strong of iterations and turns rate in reaching
optimisation ability, in path planning, best-known
effective in dynamic adaptive adjustments solutions, effective
environments improve performance, in large-scale
elite retention strategy problems,
enhances solution combines local
quality search operators
for improved
results
[107–109] Weaknesses May get trapped in Requires careful Performance can
local optima; tuning of parameters, degrade with
performance can vary Can be substantial
based on parameter computationally problem instances
settings intensive
[108,109,111] Hybrid Combines with other Hybridised with ACO Integrated with
approaches algorithms like PSO for improved path variable
for enhanced planning in intelligent neighbourhood
performance vehicles descent for better
optimisation
(Continued)
3656 Comput Mater Contin. 2025;82(3)

Table 3 (continued)

Reference Aspect Swarm intelligence Genetic algorithm Ant colony

optimisation
[113,114] Scalability Scales well with Scalable but with Scalable but faces
decentralised control, increased computational
suited for multi-agent computational costs limits with large
systems for larger populations path networks
[108,109,111] Parameter Moderate; sensitive to High mutation, High; dependent
sensitivity swarm size, inertia, crossover, and on pheromone
and social factors selection rates require decay, deposit rate,
fine-tuning and ant population
size

3.2.5 Statistical Learning

Apart from machine learning, statistical learning can be defined as mathematics intensive, which
means formalised relationships among the variables in a mathematical equation. Also, it forms a hypothesis
before building the model; besides, it depends on the smaller dataset with some attributes and operates
under assumptions compared to ML (because ML requires a massive data set to learn algorithms and is
not assumptions dependent) [39]. However, it could be categorised into three categories: natural language
processing, gesture recognition, and speech recognition (Fig. 3). Firstly, NLP is the AI approach that may
help the system or machine recognise, manipulate, and interpret human language efficiently. Its significance
in the automotive field is massive in utilising structured and unstructured data for various cases, such as
driving improvement, customer service enhancement, and resource management improvement [115]. Also,
this technique could assist the driver at the time of the journey if the driver asks for any assistance regarding
the vehicle owner’s manual, where it can convert the information of the manual into text [116]. Secondly,
gesture recognition is the approach to establishing human-machine interaction using only body actions
instead of voice. In the automotive field, gesture recognition is usually conducted through three stages,
i.e., detection, tracking, and recognition. Tateno et al. [117] developed a gesture-based control system using
an infrared sensor array installed in vehicles. The system recognises seven distinct hand movements to
control the vehicle’s directional movement through a CNN-powered data processing mechanism. Finally,
speech recognition involves understanding a person’s words through a computer and translating them into a
machine-readable format. This AI feature facilitates driving vehicles without distraction to ensure a safe and
comfortable journey in the vehicle. Its application for different purposes in automotive is also notable, and
employment of this technique is escalating with changing times. For example, a DL-based speech recognition
system was developed for Arabic drivers to reduce distraction and improve human-vehicle interaction [118].
The study achieved high speech recognition accuracy of over 94% for noise-free audio and over 91% for noisy
audio using various acoustic models. However, implementing the above approaches in the automotive sector
presents numerous challenges. NLP systems must handle context understanding, multilingual support, and
integration with other systems. On the contrary, gesture recognition faces accuracy, real-time processing, and
user variability issues. Furthermore, speech recognition systems must overcome noise interference and low
signal-to-noise ratio conditions, with advanced speech enhancement techniques offering potential solutions.
Comput Mater Contin. 2025;82(3) 3657

Voice Assistant: Contextual

responses and suggestions
Emergency Communications: Adjusting based on driver request
vehicle settings and querying vehicle
information through voice commands Translation Tool: Translating
Natural Language road signs or assisting passengers
Processing in multilingual environments
Voice Activation: Enables
hands-free control for phone, Touchless Control: Controlling
in-car functions like music,
calls, music, navigation, and
emergency responses climate, and navigation by using
hand gesture
Personalization:
Safety Alerts: Detects driver
Recognizes drivers' voice to
alertness or fatigue through
adjust personalized settings
Speech Gesture head and hand movements
(seat position, temperature)
Recognition Recognition Passenger Control: Adjust
Real-Time Transcriptions: Autonomous Vehicle
lighting or seat position
Transcribes spoken instructions
or messages without looking
away from the road

Figure 3: Automotive applications of statistical learning approaches

3.2.6 Game-Theoretic Learning

Game-theoretic learning is an approach in the AI field where a player among a set of players as
the individual agent can make the optimal decision (prediction) of each other sequentially for upcoming
actions, and the outcome comes from these decisions [119]. So, it explains how the agents strategically might
learn from the consequences of social interaction in the long run and adopt the behaviour. It has shown
significant promise for self-driving vehicles in enhancing decision-making processes and improving safety
and efficiency in complex traffic environments. Various studies have integrated game theory with deep
reinforcement learning to model interactions between multiple vehicles, including human-driven ones, in
diverse scenarios. For instance, one approach combines DRL with game-theoretic decision-making to enable
AVs to navigate unsignalized intersections by modelling different driving behaviours and anticipating the
reactions of other vehicles without explicit communication [120]. Another study formulates lane-changing
decisions as games between pairs of vehicles, using neural learning to adjust payoff matrices and optimise
behaviour [121]. Additionally, game-theoretic planners have been developed for competitive scenarios,
such as autonomous racing, where vehicles employ strategies like blocking and overtaking to outperform
others. These planners often outperform traditional model-predictive control methods by considering
vehicle interactions [122]. Furthermore, multi-agent planning algorithms have been proposed to handle
interactions with human drivers at intersections and roundabouts, assigning priorities based on observed
driving styles to prevent collisions and deadlocks [123]. However, multi-agent interaction involves predicting
and responding to the behaviours of numerous traffic participants, including other AVs, human-driven cars,
and pedestrians. This complexity is compounded by the need for real-time decision-making, which requires
efficient computational algorithms to ensure safety and optimal performance. The integration of game theory
with DL and memory neural networks has also been explored to predict opponent behaviours and enhance
path planning in dynamic environments [124]. However, robustness against unpredictable behaviours of
human drivers and the dynamic nature of traffic environments also poses a significant hurdle, necessitating
advanced models that can adapt to varying conditions and uncertainties.

3.2.7 Fuzzy Logic

Fuzzy logic is a reasoning method in AI that imitates human reasoning for decision-making by
manipulating or representing uncertain information based on the degree of truth or a partial truth that could
be all intermediate possibilities for any actual number between o to 1 [125]. Vehicle designers have leveraged
fuzzy techniques in the automotive sector to enhance dynamic control analysis, providing a cost-effective
3658 Comput Mater Contin. 2025;82(3)

solution for examining vehicle performance during driving modelling. One commonality of this method for
application in different automotive aspects is its ability to handle uncertainty and subjectivity, which is crucial
in applications ranging from failure mode and effects analysis (FMEA) to demand forecasting and energy
efficiency evaluation [126,127]. For instance, fuzzy logic enhances traditional FMEA by incorporating expert
knowledge to more effectively analyse subjective failures [128], and it is also used to predict vehicle ranges in
EVs by considering dynamic parameters [129]. Moreover, its applications extend to Industry 4.0 readiness
assessments, demonstrating versatility across different analytical domains [130]. These applications highlight
the flexibility of fuzzy logic but also underscore the differences in its implementation based on the problem
being addressed. Contradictions arise in the perceived efficiency and accuracy of fuzzy logic models. For
instance, while the Mamdani-type fuzzy logic model is praised for its high accuracy in evaluating vehicle
energy efficiency [131], other studies suggest that integrating fuzzy logic with different techniques, such as
neural networks or big data analytics, can yield more robust and reliable results [126,132]. This denial suggests
that fuzzy logic is a powerful tool. Still, its effectiveness can be context-dependent, and combining it with
other methodologies may enhance its utility in the automotive sector.

3.2.8 Hybrid Intelligent Systems

In the automotive sector, hybridisation systems combine multiple AI methodologies, such as machine
learning, rule-based systems, and optimisation algorithms, to enhance decision-making and control systems,
particularly in autonomous driving, predictive maintenance, and driver assistance systems. These approaches
share an analogy in their objective to improve accuracy, adaptability, and robustness by compensating for
individual limitations of each AI technique, such as integrating rule-based systems for interpretability with
ML for adaptability. For instance, the use of hybrid human-AI in semi-autonomous driving systems aims to
leverage human judgment alongside AI capabilities to improve system reliability and performance [133]. Sim-
ilarly, the ML-multi-agent system framework combines ML models with rule-based agents from symbolic AI
to address the limitations of ML-only approaches, thereby enhancing the driving score in fully autonomous
vehicles [134]. Recently, Rahim et al. [135] developed a CNN-BiGRU-based hybrid decision model for
vehicular engine health monitoring that combines CNN for feature extraction and BiGRU for handling
sequential dependencies in time-series engine data. They achieved high decision accuracy (0.8897) with low
decision loss, surpassing conventional models by integrating sensor data with structural and vulnerability
assessment information for improved PdM. Furthermore, hybrid systems offer a promising approach to
overcoming some of the current limitations faced by fully autonomous vehicles. By integrating hybrid electric
vehicle technology with autonomous driving systems, several practical benefits can be realised.

Energy Efficiency and Management

Hybrid systems can significantly enhance the energy efficiency of AVs. For instance, the integration of
intelligent driver assistance systems, such as adaptive cruise control (ACC) and energy management systems,
can optimise energy consumption. These systems use advanced control strategies like model predictive
control and neuro-fuzzy systems to maintain safe distances and manage acceleration, thereby reducing
energy waste and improving fuel efficiency by up to 2.6% [136]. Additionally, hybrid powertrains allow AVs
to achieve similar energy savings with smaller motors compared to those required by human drivers, thanks
to more optimal torque requests and efficient engine operations [137].
Comput Mater Contin. 2025;82(3) 3659

Enhanced Control and Safety

Hybrid control models can improve the robustness and adaptability of AVs under various driving con-
ditions. For example, combining model predictive and Stanley-based controllers can enhance vehicle control
and path-following capabilities, ensuring better performance in diverse scenarios [138]. Moreover, hybrid
ML-based control strategies can optimise driving performance and smoothness, addressing limitations in
existing control systems [139].

Environmental Impact
Hybrid systems in AVs can also contribute to reducing environmental impact. By employing advanced
fuzzy logic control for energy management, hybrid electric AVs can decrease fuel usage and extend battery
life, leading to a reduced carbon footprint [140]. This aligns with the automotive industry’s goals of
sustainability and reduced emissions.

Integration with Urban Traffic Systems

In urban environments, hybrid systems can facilitate the coexistence of autonomous and non-
autonomous vehicles. For instance, creating exclusive lanes for AVs and allowing them to travel in platoons
can reduce travel times and improve traffic flow, as demonstrated in simulations using real-world data [141].
This hybrid approach can ease the transition to fully autonomous urban mobility.
By enhancing energy efficiency, control, safety, and environmental sustainability, and by facilitating
integration with existing traffic systems, hybrid systems can address many of the current limitations of
AVs. However, the development and deployment of these hybrid systems face complexities related to
algorithm portability and the standardisation of software development processes, which are crucial for
mass production and industrial applications [142]. Subsequently, integrating AI into existing vehicle systems
requires significant modifications and validation to ensure safety and compliance with regulatory standards,
which can be resource-intensive and time-consuming [143].

3.2.9 Computer Vision

In AI, computer vision for vehicles empowers them to perceive everything in their surroundings or natu-
ral environment without human assistance. The application of computer vision techniques in the automotive
sector has been explored extensively, revealing converging and diverging viewpoints among researchers.
Several studies highlight the potential of computer vision for enhancing safety and efficiency in automotive
applications. For instance, edge and ellipse detection, camera calibration, 3-D reconstruction, and stereo
vision have been successfully applied to detect rims, estimate calibration angles, and reconstruct vehicle
trajectories, demonstrating the versatility of these methods in solving complex vehicle problems [144].
Similarly, integrating this method with Industry 4.0 technologies, such as remote sensing data fusion and
mapping tools, has improved collision avoidance and environment mapping [145]. However, there are notable
differences in focus. At the same time, some researchers emphasise the role of computer vision in ADAS
for safety purposes, highlighting the superiority of DL techniques over traditional methods [146]. Others
concentrate on its application in manufacturing for quality control and defect detection [147,148]. While
some studies advocate for the immediate integration of self-adjusting computer vision systems to achieve
zero-defect manufacturing [147], others point out the slow-growing R&D infrastructure in certain regions,
which could hinder such advancements [145]. Additionally, the use of computer vision for non-contact
weigh-in-motion systems [149] and error detection in vehicle parts [150] further illustrates the diverse
applications and ongoing research gaps in this field. However, Table 4 compares various criteria of the AI
methods mentioned above.
 Table 4: Comparison of prominently employed AI in the automotive sector
3660

AI method

Criteria Machine learning Deep learning Heuristic Meta-heuristic Statistical learning Game theory Fuzzy logic Hybrid systems Computer vision
Reference [151,152] [153,154] [155,156] [157,158] [116,159] [122,160] [161,162] [163,164] [145,153]
Data dependency High Very High Low Low High Medium Medium Variable High
Complexity Medium High Low Medium Medium High Medium High High
√ √ √ √ √
Interpretability Low Medium Variable Low
√ √ √ √ √
Real-time Limited Medium Limited Variable
feasibility
√ √ √ √ √ √ √
Adaptability Low Medium
√
Use of neural Optional × × Optional × × Optional Optional
networks
√ √ √
Optimisation Medium High Low Medium Low Medium
focus
√ √ √ √
Uncertainty Limited Limited × × Limited
handling
Common Prediction, Image, speech Optimisation, Optimisation Prediction, Decision-making Control systems Multimodal Visual
applications classification recognition routing regression problems recognition
√ √ √ √ √
Data-driven × × × ×
approach
√
Note. —Indicates suitability or presence of the characteristic. ×—Indicates the characteristic is generally not applicable or utilised. Variable—Indicates that it
depends on specific implementations or use cases . Limited—Some use cases may not fully support this characteristic.
Comput Mater Contin. 2025;82(3)
Comput Mater Contin. 2025;82(3) 3661

3.2.10 Comparative Analysis

Various AI algorithms in the automotive sector have been critically evaluated and compared based on
key performance metrics, as shown in Table 5.

3.3 Overview of AI’s Applications and Opportunities

Recently, AI’s application in the automotive sector has been increasing outstandingly to ensure intelli-
gent services, computing, networking, and security purposes, exploiting the most efficient techniques. The
prominent purposes of AI in this sector are described as follows.

3.3.1 Advanced Driver Assistance Systems

Advanced driver assistance systems are a recent automotive innovation that enhances safety and
comfort by automatically detecting and responding to potential road hazards through environmental
sensing and interpretation [177]. ADAS provides automobiles with a mix of sensor technologies and AI
processing algorithms that perceive the environment surrounding the vehicle, process it, and then either
present information to the driver or take action [178]. The state-of-the-art sensors used in this system for
various functions are RADAR (radio detection and ranging) (long range, short or medium range), LiDAR
(light detection and ranging), camera (monocular, stereo, and infrared), and ultrasonic sensors [179]. The
integration of AI in ADAS has led to the development of sophisticated algorithms, which enable the system to
adapt to different driving environments and user preferences [180,181]. To assure the highest degree of safety,
reinforcement learning algorithms have recently been applied for many levels of task repetition. For example,
Gonjalo et al. [182] created a prototype variable message sign reading system with the help of ML techniques
to improve road safety, in which images were preprocessed and delivered to an optical character recognition
model, and subsequently converted to announcements using IBM Watson Text to Speech. Moreover, Salve
et al. [183] proposed a CNN model with two output layers that can provide an alert through an alarm when the
driver is continuously predicted to be exhausted. The applications of ADAS consist of adaptive cruise control
(ACC), collision avoidance (CA), and automatic parking, among many others [184], which are discussed
as follows.

Adaptive Cruise Control

This critical ADAS component integrates onboard sensor data to maintain safe following distances and
prevent forward collisions with preceding vehicles [185]. RADAR technology emits microwave sequences
to estimate object distance and speed, while LiDAR provides high-resolution object detection by projecting
beams and calculating distances in ACC-equipped vehicles [186]. AI techniques, particularly DL neural
networks, have been employed to improve the predictive control of ACC systems. For instance, neural
network predictive control has been utilised to mimic vehicle dynamics and optimise control actions,
demonstrating improved performance in maintaining safe distances and adapting to dynamic driving
conditions [187,188]. Moreover, personalised ACC systems, which adapt to individual driving styles using
AI, have also been developed to enhance user comfort and safety by learning and adjusting to real-time
driver feedback [189]. Furthermore, innovative methods for scene recognition and target tracking in complex
traffic environments have been proposed, leveraging multi-sensor fusion and advanced filtering techniques
to improve the robustness and accuracy of ACC systems [190]. According to an analysis, traffic congestion
could be efficiently eliminated when 25% of highway vehicles are equipped with ACC [191].
3662 Comput Mater Contin. 2025;82(3)

Table 5: Performance comparison of various AI algorithms in automotive applications

AI techniques Algorithm Performance analysis in automotive applications Reference

Machine learning SVM For vehicle detection in crowded traffic, SVM [165,166]
demonstrated superior accuracy, achieving a
performance of 95.85% compared to KNN’s 92.65%.
Additionally, SVM has been used in predictive
maintenance for electrical systems, where it achieved
a balanced performance with an accuracy of 85%.
KNN In the context of parking space availability prediction, [166,167]
KNN, along with DT and RF, outperformed more
complex algorithms like multilayer perceptron,
suggesting its effectiveness in less complex prediction
tasks. In predictive maintenance for electrical
systems, KNN showed lower performance with an
accuracy of 78%, indicating that it might not be the
best choice for such applications.
RF In predictive maintenance for electrical systems, RF [165,166]
achieved the highest classification accuracy of 92%
and showed robustness in regression tasks with the
lowest mean absolute error. Additionally, in parking
space availability prediction, RF was among the top
performers, indicating its effectiveness in handling
complex datasets and providing reliable predictions.
Deep learning CNN CNNs are instrumental in AVs for tasks such as [168,169]
object detection, lane detection, and traffic sign
recognition. These are also used in predictive
maintenance to detect anomalies in visual data from
vehicle components. CNN achieves the highest
accuracy (83.45%) in road image recognition for
AVs, outperforming other algorithms like SVM and
RF in classifying images into categories such as left,
right, forward, and stop.
LSTM LSTMs are used to predict future states of vehicle [170]
components based on historical data, which is crucial
for scheduling maintenance and preventing
breakdowns. They outperform traditional methods in
terms of accuracy and efficiency, as demonstrated in
predictive maintenance models where LSTM-GANs
achieved high fault prediction accuracy.
DRL DRL algorithms, such as DQN and proximal policy [93,171]
optimisation, have shown significant promise in
improving the performance of AVs by enabling them
to learn complex behaviours through interaction
with the environment. It is also applied in routing
algorithms for automotive supply chains to improve
efficiency and adaptability.
Heuristic Flexi algorithm In the context of software integration, the algorithm [172]
has been shown to efficiently schedule control
applications on shared resources, minimising control
performance degradation. This approach is
particularly useful when dealing with the integration
of multiple distributed periodic control tasks, as it
provides nearly optimal results with significantly
reduced computation times compared to more
exhaustive methods like constraint programming.
Meta-heuristic GA and PSO Widely used for their robustness and efficiency in [107–109]
control systems, predictive maintenance, and supply
chain management. Each algorithm has its strengths
and is chosen based on the specific requirements of
the application, such as cost efficiency, convergence
speed, and accuracy.
(Continued)
Comput Mater Contin. 2025;82(3) 3663

Table 5 (continued)

AI techniques Algorithm Performance analysis in automotive applications Reference

Game theory N/A Game-theoretic approaches have been used to [173,174]
develop robust path tracking control systems for
AVs, ensuring superior tracking performance under
varying conditions. Additionally, game theory has
been applied to optimise traffic flow at urban
intersections, where it helps in resolving conflicts
between vehicles and improving overall traffic
efficiency.
Fuzzy logic N/A In hybrid EVs, fuzzy logic controllers are used to [175,176]
optimise motor speed and throttle control,
enhancing vehicle performance and reducing
emissions. The effectiveness of these controllers is
often improved through self-tuning using heuristic
algorithms, which adjust the gain factors for better
performance. Furthermore, fuzzy logic is integrated
with meta-heuristics in simheuristics to tackle
complex transportation optimisation problems,
allowing for the modeling of uncertainties in travel
and service times.

Collision Avoidance
Collision avoidance, a sophisticated ADAS component in autonomous vehicles, prevents potential
collisions by alerting drivers to environmental hazards [192]. It usually follows two strategies for risk
assessment, i.e., path planning and path tracking strategies [191], and the process is briefly presented in Fig. 4.
The path planning component designs a collision-free route by analysing current road conditions. In contrast,
path tracking ensures the vehicle adheres to this route using feedback adjustments. The risk assessment
is used to quantitatively analyse the objects’ risk level from where collisions can be predicted. After a
proper risk assessment of imminent collision, the vehicle will replace the current route and follow the
replanned route using a path-tracking strategy. It addresses various driving scenarios, such as lane-keeping
and lane-changing, by configuring control parameters specific to each scenario. The system relies on vehicle
actuators—steering, throttle, and braking—to execute manoeuvres and adapt in real time. Two controllers,
longitudinal (for speed control in car-following) and lateral (for lateral movement in lane-changing)
work together to maintain safe and stable vehicle behaviour. Yu et al. [193] developed spatial-temporal
CA algorithms using ML to enhance collision prediction accuracy and optimise AV decision-making.
Additionally, a CA model was proposed by Rill et al. [192] that used optical flow and monocular depth
estimation based on DL techniques to estimate the ego vehicle’s speed depending on the lead vehicle’s
distance. However, challenges like restricted fields of view and inaccurate range information generated from
vision-based methods have been addressed by developing a multi-sensor approach that integrates LiDAR
data for precise time-to-collision calculations [194].

Automatic Parking
Automatic parking systems represent a pinnacle of vehicle automation, seamlessly integrating real-time
sensor data with sophisticated motion planning algorithms. By synthesising environmental inputs from an
array of sensors, these systems can execute precise parallel, angled, or perpendicular parking manoeuvres
autonomously, effectively removing the complexity and stress traditionally associated with challenging
parking scenarios [191]. Ultrasonic sensors are used in these systems to store information about parking
spaces when a vehicle exits its parking spot. Afterwards, based on a two-dimensional map, RADARs are
3664 Comput Mater Contin. 2025;82(3)

employed to locate appropriate parking places [195]. Recently, numerous researchers have applied DRL
to develop an automatic parking model to ensure more complex and efficient operations. For example,
Zhang and coworkers [196] proposed that reinforcement learning-based end-to-end parking algorithms
allow vehicles to learn optimal steering commands through continuous experience, thereby reducing path-
tracking errors and improving parking accuracy. Additionally, the fusion of camera sensor data and RADAR
enhances these systems’ perception and decision-making capabilities, enabling more reliable multi-target
matching and scenario extraction [197]. Furthermore, Guo et al. [198] designed an automatic parking system
using an improved DL algorithm, which resulted in 34.83% less time-consuming than manual parking.

Path Planning

Driving Scenarios

Longitudinal Controller
Collision Avoidance
Control System Reference Speed Tracking
i.e. Car-Following, Lane-Keeping

Control
Path Tracking Methodology
Achieved by Vehicle Actuators Lateral Controller
(Steering, Throttle, and Braking)
Reference Path Tracking
i.e. Lane-Change

Actuation
Configuration

Figure 4: Collision avoidance system using AI

3.3.2 Vehicle Health Monitoring Systems

AI-enabled vehicle health monitoring systems (VHMS) use sensor networks and diagnostic algorithms
to monitor component conditions and detect abnormalities in modern vehicles. The system collects infor-
mation from the engine body, fuel flow system, engine speed, etc., using connected sensors, Arduino, and
the cloud [199]. Therefore, any abnormal behaviour from single or multiple vehicle components is reported
instantly through VHMS so that safety precautions can be taken rapidly to avoid complete vehicle failure.
Recent studies have emphasised the integration of AI technologies to enhance vehicle health management
capabilities. For instance, Rahman et al. [200] have suggested a conceptual framework for a central VHMS
based on multi-layer heterogeneous networks. They advocated an ML approach in this scenario to monitor
individual vehicle health concerns, warn drivers, and create data for future necessary measures. Moreover,
the use of DL for monitoring engine vulnerable components [201], signal integrity tracking, and fault
detection [202] are highlighted as some of the significant advancements. Similarly, the application of AI
for real-time diagnosis and prognosis in vehicular systems is underscored, with a focus on the powertrain
Comput Mater Contin. 2025;82(3) 3665

and drivetrain components. However, there are notable differences in the approaches and technologies
employed. Though some researchers advocate using unsupervised DL models to simplify data collection
and labelling processes [202], others propose using dynamic Bayesian networks for fault detection and
prognosis in AVs [203]. Additionally, using lightweight computational intelligence for IoT health monitoring
in off-road vehicles, leveraging economic sensors and edge devices presents a cost-effective alternative to
traditional methods [204]. Nevertheless, researchers contrast the dependence on cloud-based vs. edge-based
solutions, with some studies emphasising the need for cloud computing to handle big data [202]. In contrast,
others highlight the limitations of cloud dependence in rural areas and propose edge-device-enabled
solutions [204]. These perspectives underscore the diverse methodologies and technological preferences in
AI-based vehicle health monitoring, reflecting both the potential and the challenges of implementing such
systems across different vehicle types and operational environments.

3.3.3 AI in Automotive Manufacturing

The integration of AI in automotive manufacturing spans across design and prototyping, manufacturing
and assembly, and post-production systems, offering numerous advantages as illustrated in Fig. 5. For
example, General Motors in the USA employed AI to create a prototype for one of their vehicles that
evolved a new seat belt bracket that was 40% lighter and 20% stronger than its original design [205].
In addition, to improve the flexibility of manufacturing in the automotive metalworking industry, a low-
load universal collaborative robot (cobot) system was proposed with innovative security solutions [206].
AI technologies have significantly enhanced design efficiency, reduced costs, and improved manufacturing
quality, making the industry more competitive and sustainable [207]. For instance, using federated learning-
AI in decision-making and smart contract policies effectively managed costs, energy, and other control
functions in the manufacturing process [208]. Moreover, the automotive AI transformation has radically
changed organisational infrastructure and management working methods for better performance, efficiency,
and competitiveness. At the same time, leadership styles are also evolving in automotive industries, and
AI methods are being used to provide more strategic and analytical solutions within the optimum time
frame [209]. Workforce adaptation and ethical considerations are significant impediments, as the shift
towards AI-driven processes necessitates new skills and raises concerns about job displacement [207].
Moreover, it is driving innovation and reshaping the industry’s future. There are difficulties in establishing
AI technologies at the firm level, emphasising the need for organisational readiness and digital skills [210].
Contradictions also arise regarding the scope of AI’s impact. Some researchers focus on the benefits of AI in
autonomous vehicle design and production logistics [211]. In contrast, others stress the unresolved issues in
production planning and sequence control [212].

3.3.4 AI in Supply Chain Management

Supply chain management (SCM) is the process of controlling the movements of products and services,
which covers all procedures that convert raw materials into finished commodities. Different AI technologies
are currently being used in various subfields of automotive SCM for resource utilisation and information
sharing among suppliers, manufacturers, and consumers to get the highest efficiency keeping time and
money at an optimum level (Fig. 6). As an example, study by Arunmozhi et al. [213] found a 12.48%
reduction in energy wastage and an 11.58% reduction in hidden financial transactions in the AV supply
chain using the blockchain and AI-based approach. Their framework is expected to improve product
traceability, transaction transparency, and sustainable economic growth in AV supply chains. Subsequently,
AI-based SCM significantly influences raw material procurement, bridging the gap between raw material
suppliers and automobile manufacturers [214]. AI and ML effectively maintain a proper and accurate
3666 Comput Mater Contin. 2025;82(3)

inventory of automotive parts by analysing large real-time data sets and forecasting supply and demand
efficiently to improve product availability and customer satisfaction [215]. AI can optimise inbound logistics
in the automotive industry by assessing disruption risks and proposing countermeasures to ensure material
availability [216]. AI techniques combined with IoT help track and monitor automotive parts to avoid
any unprecedented damage (due to improper handling and by providing insights into materials based
on temperature and humidity) and ensure timely arrival at sites [217]. Furthermore, in the production
process, AI is able to identify any quality issues of the finished product at an early stage. For instance, Audi
implemented a project to improve the testing process in products where ML automatically identified the
finest cracks of sheet metal parts within seconds [217]. Meanwhile, Mercedes Benz and Renaults came to
realise the benefits of AI in the supply chain as soon as they developed risk mitigation and contingency plans
to reduce the probability of production or financial losses where large amounts of data were processed with
the help of AI models [205].

High-Performance Computing

Computer-Aided Design Models

Simulations for Prototyping

ML for Design Optimization

Prototype
Design
Benefits: Faster Design Approval,
Enhanced Workflow

AI-based Automotive Workflow Management and

Predictive Maintenance Alerts
Manufacturing Scheduling
Defect Identification before Safety-Enhanced Robots Working
Assembly Alongside Human
Smart Picking & Sequencing for
Quality Inspection after Assembly
Efficiency
Post- Continuous Learning & Manufacturing Radio Frequency Identification for
Production Improvement based on Data & Assembly Inventory Control
Benefits: Higher Quality Control, Benefits: Cost Reduction, Reduced
Minimize Risks, and Efficient Downtime, and Increased
Maintenance Productivity

Figure 5: AI deployment in automotive manufacturing

3.3.5 Predictive Maintenance

Predictive maintenance intends to predict the time frame of maintenance activities based on the system’s
health condition (statistical PdM) and recorded data of the system (condition-based PdM) to avoid costly
repair as well as ensure timely repair (before failure). Fault diagnostics and prognosis are important research
areas for PdM involving feature extraction, selection, and fault classification. By leveraging advanced sensor
technologies and big data analytics, AI techniques can predict potential failures and estimate the remaining
useful life (RUL) of automotive components, allowing for timely and precise maintenance actions [218–
220]. This proactive approach minimises unexpected breakdowns and extends the lifespan of vehicle parts,
Comput Mater Contin. 2025;82(3) 3667

contributing to overall operational efficiency [221]. Table 6 shows that most existing research relies on
supervised ML methods requiring labelled data. Therefore, combining multiple data sources can improve the
accuracy of PdM models [222]. Modern vehicles equipped with sensors provide extensive data about their
condition and performance, which can be analysed to avoid vehicle failures by constructing statistical and
mathematical models using ML algorithms and executing maintenance activities resulting in models [218].
For instance, Aravind et al. [223] presented a model for AI-based PdM in AVs using physics models,
advanced sensors, and reconfigurable devices, aiming to examine electronic components’ deterioration and
the relationship between local failures and global system malfunctions. The research presents a proof-of-
concept simulation that demonstrates the effectiveness of the proposed physics models, data-driven learning
strategies, and virtual prototype-aided dynamics learning control schemes in handling different transmission
system dynamics and fault phenomena. However, the non-availability of real automotive datasets, usually
considered highly confidential by companies, prevents a qualitative comparison of novel approaches with
the state of the art. Furthermore, evaluating the validity of developed methods using real data is difficult, as
real data is often unlabelled or only partially labelled, and annotating data is time-consuming and requires
expert knowledge [218].

Inventory Management
& Demand Forecasting

Planning &
Finance
Operations

Resource Production &

Management Manufacturing
Artificial Intelligence
in SCM

Intelligent Risk Logistics &

Management Transportation

Customer Service
(Sales & After Sales)

Figure 6: Features of AI-driven automotive SCM

3668 Comput Mater Contin. 2025;82(3)

Table 6: AI techniques used in PdM of automotive cases

Reference Features Predict cases AI techniques

Narayanan et al. [224] Engine temperature, oil Armored fighting ML algorithms
pressure, fuel vehicle engine failures
consumption, rpm, and
vibration levels
Wang et al. [225] Vehicle maintenance Vehicle maintenance Seq2Seq neural
time series data prediction network with GRU
Sharma et al. [226] Historical data on RUL of tyres DT, KNN, RF
vehicle tyres
Chen et al. [227] (GIS) data (weather, RUL Cox proportional
traffic, terrain) hazard model,
Merged-LSTM
network
Vasavi et al. [228] Engine load, rpm, Vehicle health ANN + KNN
pressure sensor, air monitoring
flow sensor, engine
coolant temperature
sensor, throttle
position, air intake
temperature, and fuel
level
Biddle et al. [229] Driving torque, AV’s controller fault SVM
steering wheel angle, detection and
and brake pressure prediction
Hu et al. [230] Vibration signals and Suspension health LSTM
fatigue loads monitoring and
remaining fatigue life
Mckinley et al. [231] NOx sensor data NOx sensor failure Xtra gradient boosts
(engine out NOx , ML
system out NOx )
Krishnamurthy Imaging-based sensors Component wear Feedforward neural
et al. [232] network
Voronov et al. [233] Sparse vehicle Battery failure LSTM, RF
operational data
Chen et al. [89] GIS data (climatic Time-between-failures DNN, RF, KNN
features) of vehicle
Wang et al. [234] Vehicle age, early Durability of Conditional inference
failure, and repair automobile engine tree
history
Shafi et al. [235] Sensor data from fuel, Failure prediction in DT, SVM, KNN, RF
ignition, exhaust, and vehicle subsystems
cooling systems
(Continued)
Comput Mater Contin. 2025;82(3) 3669

Table 6 (continued)

Reference Features Predict cases AI techniques

Giobergia et al. [236] Time series data from Oxygen sensor Custom feature
the engine control unit clogging selection algorithms,
semi-supervised
labelling, and
classification
algorithms

4 Conceptual Framework of AI-Based Electric Vehicle’s Real-Time Health Control Model

The significance of AI in the automotive sector in regulating the degradation of vehicles and their control
systems’ valuable parts is growing because of the real-time monitoring of their health and performance. Until
today, the available intelligent features in traditional vehicles successfully minimised maintenance expenses,
improved durability, reduced unexpected downtime, unexpected accidents, and so forth [237]. However,
conventional vehicles are not environmentally friendly due to using fossil fuels to drive them, the emission of
excessive greenhouse gasses, and climate deterioration. As a result, the demand for electric vehicles is rising
dramatically, and the market analyst predicted that the market will expand to $802.81 billion by 2027 [238].
Again, according to Reuters, vehicle manufacturers commit to ending fossil-fuel vehicles by 2040 with 100%
zero-emission vehicles. Based on the discussed matters, it can be inferred that EV demand will explode
quickly instead of fossil-fuel-driven vehicles because EV production costs will be cheaper and more flexible
than fossil fuel vehicles by 2027, and the fossil fuel vehicles’ overall market could be downed [239,240].
The notable matter is that EVs have many essential components, i.e., complete structure, electric
controller, electric motor, tyre, transmission system, and so on, which are greatly responsible for running
the vehicle smoothly. Unexpected irregularities of these components may lead to devastating accidents,
reduce vehicular longevity, increase operating expenses, waste valuable time, and many other issues may
arise [241]. Also, sensible component irregularities may lead to unexpected downtime, irregular spare parts
replacement, excessive maintenance expenses, excessive operating expenses, and replacement of expensive
spare parts, hindering people’s interest in EVs. Recently, much research has been conducted on EVs as part of
their health monitoring system to overcome the abovementioned issues. Still, these researches are confined
to specific items of health monitoring systems like tyres [242], batteries, traction motors [243], charging
faults [244], predicting the RUL of EVs [245], and so forth. Therefore, in the Industry 4.0 era, the smart and
intelligent decision-providing model is a prerequisite to control the degradation of EV’s sensible component
by measuring its real-time condition. However, till now, a robust, smart, intelligent, and secure AI-based
EV’s real-time health control model system has not been developed using structural data and associated
risk factors that could monitor, control, and notify as fast as possible by ensuring reliability and accuracy.
Therefore, Fig. 7 presents an AI-based comprehensive model designed to monitor, assess, and manage the
health of critical EV components in real time. The process begins with identifying essential components like
the battery, motor, and transmission and then assessing their vulnerabilities and associated risk factors. A
wireless sensor network is established to continuously gather real-time data from these components, focusing
on key indicators like temperature, vibration, and electrical parameters. The collected data undergoes
preprocessing and feature engineering to ensure its quality and relevance for predictive analytics. ML and
DL algorithms then analyse this data to estimate each component’s RUL and health status, enabling accurate
predictions of potential failures. After rigorous model performance evaluations in a controlled test-bed
3670 Comput Mater Contin. 2025;82(3)

environment, the VHMS framework is implemented in real-world EVs applications, where it continuously
monitors component health and provides timely alerts and maintenance recommendations. This smart,
intelligent approach minimises unexpected downtime and maintenance costs and supports the Industry 4.0
goal of creating smart, sustainable, and efficient EVs.

Finalize Research Target Through

Preparing for Research Target Through
Collecting Significant Information about Probable
Vehicular Essential Elements Sensible Components of Electric Vehicle (EV)

Vulnerability Assessment Identify EV's Sensible Components using

Identification Framework
Decision Model System
Identify EV's Risk Factors for the Decision Model

Ensure Targeted Goals Efficacy and Release for

Real-World Application Through
Making Real Targeted Goals Through
Data Generation, Data Preprocessing, Labeling,
Feature Engineering Development of Wireless Sensor Network (WSN)
among Significant Components using Zigbee
Data Feeding to ML/DL Algorithms
Development of EV's Real-Time Health
Evaluate Performance of Developed Model Monitoring and Control Model

Test-Bed Implementation of AI-Enabled Model

Recommend for Real-World Application

Figure 7: Conceptual framework of AI-based health monitoring system for EVs

5 Challenges and Future Scopes of AI in Automotive Sector

As AI technologies evolve, they promise unprecedented advancements in autonomous driving, predic-
tive maintenance, and intelligent transportation systems while presenting complex technological, ethical, and
infrastructure challenges. However, successful integration requires addressing critical issues and overcoming
many obstacles, as presented in Fig. 8.
Comput Mater Contin. 2025;82(3) 3671

Current Challenges

Calibration Reliability Workforce Adaptation Cloud-to-Edge Deployment

Real-World Transition
Multi-Sensor Fusion

Feature Engineering
Model Sphistication
High Speed Data

Ethical Concerns
Complexity &
Sensor Issues Model Tuning
Uncertainty

Biometric Authentication Transparency Privacy

Compliance Standards
Intrusion Detection

Societal Impacts
Accountability
Ethical
Cybersecurity
Challenges

Challenges & Future Directions of AI in Automotive Industry

Future Research Scopes

Deep Learning Hybrid AI Federated AI Digital Twins Cobots Ergonomics Simulation Standardization
Additive Manufacturing
Predictive Maintenance
High-Definition Maps

Continuous Learning
Adversarial Testing
Augmented Reality

Safety Measures
5G Integration

Autonomous Smart Human-Robot Testing &

Driving Manufacturing Collaboration Validation

Figure 8: Overview of AI challenges and future opportunities in automotive

5.1 Associated Challenges

Integrating AI into existing automotive systems presents several practical challenges. One of the
primary concerns is ensuring the safety and dependability of AI-based systems, especially in mission-critical
applications like real-time embedded systems in vehicles [246]. The integration process must address ethical
and legal considerations, data privacy, and liability issues, which are crucial for maintaining trust and
compliance with regulations [247]. Additionally, the complexity of AI systems requires robust solutions to
handle scalability, real-time data processing, and security vulnerabilities, which are essential for enhancing
vehicular networks and communication protocols. The automotive industry also faces challenges related to
data availability, quality, and system integration, alongside ethical and regulatory conundrums, particularly
as vehicles become more autonomous [248]. Furthermore, the integration of AI in automotive design and
manufacturing must consider the impact on employment and the need for international standards to ensure
secure AI systems in vehicles [247]. However, following are the detailed practical challenges of introducing
AI in automotive sector.
3672 Comput Mater Contin. 2025;82(3)

5.1.1 Sensor Issues

AI integration in AVs fundamentally depends on sophisticated sensor technologies, which serve as
the critical sensory interface for perception and decision-making processes. The automotive sector faces
substantial challenges in ensuring sensor reliability, performance, and effectiveness, directly impacting the
safety and functionality of AI-driven automotive systems. Sensor calibration emerges as a crucial factor
in autonomous system performance. Complex multi-sensor fusion techniques require precise alignment
for accurate object detection and reliable vehicle operation [249]. In addition, sensor reliability represents
another pivotal concern, particularly for IoT sensors in cyber-physical systems. Unreliable sensor data can
fundamentally undermine AI applications, necessitating robust methodologies to enhance data quality and
consistency [248]. Moreover, the intricate sensor infrastructure in modern vehicles demands sophisticated
data management and high-speed networking capabilities to process the vast amount of generated data. This
complexity introduces significant challenges across hardware and software domains, with additional com-
plications arising from high energy consumption requirements and the need for rapid data processing [250].
After all, innovative AI applications for the automotive sector are often limited by the sensor technology
available. The success of an entire AI model is strongly dependent on the quality of input sensor signals [27].
Hence, resolving these challenges is the first step towards the safe and effective deployment of AVs and other
AI surfer automotive applications.

5.1.2 Complexity and Uncertainty

The development of AI systems, especially for autonomous and connected vehicles, relies heavily
on the labour-intensive contributions of micro-workers who annotate and process vast amounts of data,
highlighting a structural dependency on low-paid labour across global supply chains [251]. Additionally,
the increasing complexity of vehicle digital architectures necessitates sophisticated AI models for predictive
maintenance, which, while promising enhanced efficiency and cost reduction, also reveal emerging and
declining research trends that underscore the dynamic and evolving nature of the field [218,221]. Adopting
AI in automotive safety and cybersecurity further complicates the landscape. New attack vectors and
sophisticated hacking tools pose significant risks, necessitating the development of explicable AI and
robust security measures [252]. The automotive industry’s ongoing digital transformation, driven by AI,
IoT, and ML integration, has reshaped business processes and relationships, underscoring the sector’s
embrace of innovation and technological advancement. However, this digital overhaul also presents ethical
considerations and challenges in workforce adaptation [253]. Thus, the complexity and uncertainty of AI in
the automotive sector are characterised by a confluence of technological advancements, labour dynamics,
cybersecurity threats, and ethical factors, all of which require careful navigation to harness AI’s full potential.

5.1.3 Complex Model Tuning Issues

Training complex AI models in the automotive industry presents many significant challenges of fine-
tuning. The complex and time-variant nature of system parameters in this sector presents substantial
challenges to properly defining essential features such as non-linear behaviours, time delays, and parameter
drifts to develop effective control and modelling [254]. In addition, deploying AI models in automotive
contexts necessitates advanced methodologies surrounding data collection, training, and model deployment.
Acquiring high-quality, wide-ranging datasets and building generalisable model architectures robust to
varied operating conditions make this process particularly arduous [17]. The integration of AI also increases
the complexity of the overall tuning process in terms of machine learning (for example, hyperparameter
optimisation). It introduces the need to move rapidly from cloud-based development toward edge-based
implementation to achieve adequate model performance at the required reliability rates in real-world
Comput Mater Contin. 2025;82(3) 3673

situations [255]. Furthermore, advanced feature engineering and anomaly detection methods are needed to
support defining AI models for PdM and traffic flow prediction in cooperative and autonomous EVs due
to the large amount of data produced by modern automotive systems [218,256]. Multi-objective tuning is
more complex since it is dependent on micro-workers for labelling the data, and there is a requirement to
continuously update the AI models to cater to the new data and scenes [251].

5.1.4 Cybersecurity and Functional Safety

Incorporating AI and ML in the automotive sector has dramatically increased the attack surface,
posing substantial cybersecurity risks. Adopting these technologies promises enhanced user experiences
and improved traffic management but also introduces security risks across various automotive systems,
including electronic control units, infotainment, and communication networks [257]. The complexity of
data and traffic behaviours in AV networks, controlled by protocols like CAN bus, makes them susceptible
to various attacks such as spoofing, flooding, and replaying [258]. The increasing connectivity of vehicles
exacerbates societal vulnerability to cyberattacks, necessitating robust cybersecurity validation processes and
the development of new safety integrity levels [259]. Despite the advancements, the automotive industry
faces significant challenges in knowledge sharing and collaboration among manufacturers and suppliers,
which is crucial for addressing cybersecurity threats effectively [260]. Current research highlights the
predominance of intrusion detection systems in automotive cybersecurity, yet many areas, such as biometric
authentication and cryptographic methods, remain underexplored [261]. However, lightweight crypto-
graphic protocols [262], hybrid cryptography engines [263], and hardware security modules [264] are being
developed to secure vehicular communications and components. Meanwhile, biometric solutions, including
noninvasive sensor authentication [265], and biometric fingerprinting and keystroke dynamics [261], offer
promising applications for vehicle security. As standards like ISO/SAE 21434 emphasised, comprehensive
cybersecurity testing is critical for identifying and mitigating vulnerabilities, although specific technical
guidelines are still lacking [266].

5.1.5 Ethical Challenges

The ethical challenges of AI in the automotive sector are multifaceted and complex, encompassing
issues such as human agency, technical robustness, privacy, transparency, fairness, societal well-being, and
accountability. AVs present significant ethical dilemmas, particularly in decision-making processes where
machine ethics must navigate scenarios involving potential harm to humans [267,268]. The rapid adoption
of AI technologies in automotive manufacturing and design has led to substantial improvements in efficiency
and quality. Still, it also raises ethical concerns related to workforce displacement and the need for equitable
adaptation strategies [269]. Moreover, deploying AVs necessitates robust data governance frameworks to
protect user privacy and ensure transparency in AI decision-making processes [267,270]. Despite ongoing
efforts by academia, policymakers, and industry stakeholders to address these ethical issues, a consistent and
comprehensive solution remains elusive, highlighting the need for continued collaboration and innovation
to develop ethical guidelines that can keep pace with technological advancements [268,270].

5.2 Future Research Scopes

5.2.1 Autonomous Driving and Intelligent Vehicles
One critical focus is enhancing deep learning techniques to improve autonomous driving systems’
reliability and real-time performance, particularly in complex scenarios involving road, lane, vehicle,
pedestrian, and traffic sign detection [271,272]. Additionally, integrating AI with other emerging technologies
3674 Comput Mater Contin. 2025;82(3)

such as high-definition maps, big data, high-performance computing, augmented reality, virtual reality,
and 5G communication will be essential for achieving full automation and improving the decision-making
capabilities of AVs [273,274]. The concept of hybrid human AI also presents a promising avenue, com-
bining human judgment with AI to overcome current limitations and enhance the safety and efficiency
of semi-autonomous systems [133]. Furthermore, addressing the challenges of cybersecurity, vehicle-to-
everything privacy, and risk mitigation technologies will be crucial for the safe deployment of intelligent
vehicles [274]. Finally, the development of innovative AI algorithms and hardware architectures tailored for
autonomous driving will be necessary to handle the increasing complexity and ensure the robustness of these
systems [275].

5.2.2 Smart Manufacturing

Smart contracts significantly enhance trust and efficiency in AI-driven automotive manufacturing
processes by providing a secure and automated framework for executing and controlling manufacturing
operations. In the proposed design framework [208], smart contracts are integrated with federated AI to
facilitate data-driven decision-making and process execution in a decentralised manner. This integration
introduces a trust threshold limit, which helps manage and moderate the use of resources such as equipment,
tools, energy, and costs, thereby reducing wastage in manufacturing processes. By leveraging blockchain
technology, smart contracts ensure transparency and immutability, which enhances trust among stake-
holders by securely handling trading policies and market risk assessments, even during socio-economic
crises. The model demonstrates improved decision accuracy and efficiency in procurement, assembly,
and manufacturing, ultimately optimising cost, energy, and control functions. Developing AI-based mea-
surement systems, including ADAS and intelligent industrial measuring systems, will also increase road
safety and enable PdM [25]. The application of ML algorithms for PdM is another critical area, aiming to
optimise manufacturing tool performance and reduce downtime through advanced diagnostic and fault
detection methods [276]. Furthermore, the role of AI in enhancing design efficiency, cost reduction, and
manufacturing quality underscores its transformative impact on the industry [277]. The exploration of
key enabling technologies such as digital twins, additive manufacturing, and AI-based monitoring and
inspection will also be essential for the evolution of smart factories [278]. Future research should also address
the challenges of customisation, graphical user interfaces, and cloud integration to fully realise the potential
of AI in smart manufacturing [208].

5.2.3 Human-Robot Collaboration

Developing more intuitive and seamless human-robot-environment interactions leverages humans’
cognitive capabilities and robots’ physical power to create a synergistic work environment [279]. The
deployment of AI with cobots and technologies such as augmented reality (AR) and digital twin can
significantly improve the flexibility and efficiency of manufacturing processes, addressing the pressing
needs of the automotive sector [280]. Additionally, implementing AI-enabled robots and 3D printing
technologies drives rapid growth in the automotive robotics sector, particularly with the shift towards
EVs [281]. Future research should also focus on enhancing human-robot collaboration systems’ safety
and interaction modalities, incorporating advanced manual guidance techniques and wearable devices to
facilitate multi-modal interactions in a fenceless environment [282]. Moreover, the ergonomic aspects of
head-mounted displays used in AR applications need further maturation to ensure operator comfort and
effectiveness [283]. Finally, establishing standards and guidelines for developing AR assistance systems and
conducting large-scale assessments in industrial environments will be crucial for widespread adoption and
success [283].
Comput Mater Contin. 2025;82(3) 3675

5.2.4 Testing and Validation of AI systems

As automotive AI systems become more complex, rigorous testing methodologies that mimic real-world
conditions are essential. Research should focus on creating standardised, scalable frameworks for testing
AI under diverse scenarios, including rare edge cases that current datasets may not capture. Additionally,
advancing digital twin and simulation technologies can help evaluate AI-driven automotive systems more
comprehensively while reducing physical testing costs [284]. There is also a need to explore robust validation
techniques, such as adversarial testing and continuous learning algorithms, that allow systems to adapt and
self-correct in dynamic environments. Transparency in AI decision-making remains a critical research area,
as interpretable models will enhance the industry’s ability to validate complex algorithms and make decisions
auditable and compliant with emerging automotive regulations [285]. This multifaceted research approach
will ensure that AI systems in the automotive sector are safe, adaptive, and aligned with ethical and regulatory
standards, driving forward the sector’s reliability and public adoption.

5.2.5 Integration of AI with 5G

The integration of AI with 5G in the automotive sector presents a promising avenue for future research,
focusing on enhancing vehicular intelligence, safety, and communication. As 5G technologies advance,
they offer a robust platform for AI applications, enabling more efficient vehicle-to-vehicle communication
and traffic management through mobile edge computing [286]. This integration is expected to support the
development of AVs by providing ultra-low latency and high reliability, essential for real-time decision-
making and remote driver monitoring systems [287]. Furthermore, AI can optimise 5G network operations,
such as resource allocation and network slicing, to meet the stringent quality of service requirements of
autonomous driving scenarios [288]. However, the integration also poses challenges, particularly in terms of
security threats to AI-driven systems, necessitating ongoing research into robust security frameworks [289].
As the industry moves towards 6G, AI’s role will become even more critical, potentially transforming
vehicles into highly autonomous, non-polluting, and safe devices, thus making them integral to future smart
cities [290]. This convergence of AI and 5G technologies is anticipated to create a more intelligent and
connected internet of vehicles ecosystem, driving the need for continued exploration of these synergies to
address emerging challenges and opportunities.

6 Conclusions
Artificial intelligence digitally transforms the automotive sector and can improve the vehicle’s design,
assembly and experience. This extensive study examines the varied ways in which AI shapes automotive
technology and gleaned several key insights.
• Image and sensor data processing using deep neural networks have played a significant role in
enhancing the perception capabilities of vehicles. Suitable for autonomous driving and environmental
understanding, these models are proficient in pattern recognition and can work enormously with
unstructured data.
• A key strength of reinforcement learning is that it can be used to realise adaptive vehicle systems
that make intelligent decisions in uncertain environments. However, drivers might not face the same
situation more than once, so these algorithms can simulate driving conditions and iterate until they learn
the best way to approach these real-world scenarios.
• Computer vision technology is a fundamental aspect of vehicle perception, a process of reading
visual information collected by different cameras. Such functionality supports advanced movement
operations such as lane following, pedestrian identification, and visitor signal recognition, which will
3676 Comput Mater Contin. 2025;82(3)

act as the human-like mind for on-road exercising and assist in better coordinating the vehicle on
high-speed roadways.
• The research also indicated the potential of AI in EV diagnostics, providing a practical guide for real-time
diagnostics and linking the theoretical framework to real-world implementation.
• The research notes several significant challenges, particularly the “black box” issue in complex AI
models. Concerns about accountability and trust arise from the opacity of decision-making processes.
The dependence on training data creates opportunities for bias that might hinder performance in
multiple deployment scenarios. To address the “black box” problem, several strategies can be employed.
One approach is to develop inherently interpretable models rather than relying on post hoc explanations
of complex models. This approach emphasises designing models that are transparent and understandable
from the outset, which is particularly crucial in high-stakes applications such as automotive sys-
tems [291]. Another strategy involves using probabilistic models, such as Bayesian networks, to provide
post hoc interpretations of black-box models. These models can offer insights into which features con-
tribute to predictions and why, thus enhancing the interpretability of the decision-making process [292].
Additionally, robust and stable explanation frameworks can be developed using adversarial training
to ensure that explanations remain consistent even under distribution shifts [293]. Finally, employing
techniques like Markov blankets can help focus on the most relevant features, providing a clearer
understanding of the model’s behaviour [292].
However, this study acknowledges several inherent limitations in its approach. The rapid evolution
of AI technology means that some of the most advanced features highlighted may not be fully captured.
Additionally, the study mainly relied on academic publications, which may reflect limited knowledge
originating in proprietary industrial innovations that are unable to be made public. Moreover, although the
developed conceptual framework is theoretically sound, it is more thoroughly examined in practical, real-
life applications and for many different types of vehicles and use cases. Immersing into the ever-evolving
era of Industry 4.0, this review is a solid base by which one can grasp a basic framework for where AI fits
into industry change while also exposing the yet understudied research landscapes of emerging technologies,
such as the IoT, edge computing, and hybrid artificial intelligent systems.
Acknowledgement: The authors thank Universiti Malaysia Pahang Al-Sultan Abdullah for providing laboratory
facilities.

Funding Statement: The authors are grateful to the Universiti Malaysia Pahang Al-Sultan Abdullah and the Malaysian
Ministry of Higher Education for their generous support and funding provided through University Distinguished
Research Grants (Project No. RDU223016), as well as financial assistance provided through the Fundamental Research
Grant Scheme (No. FRGS/1/2022/TK10/UMP/02/35).

Author Contributions: Study conception and design: Md Naeem Hossain, Md. Abdur Rahim, Md Mustafizur Rahman;
data collection: Md Naeem Hossain, Md. Abdur Rahim; analysis and interpretation of results: Md Naeem Hossain, Md
Mustafizur Rahman, Devarajan Ramasamy; draft manuscript preparation: Md Naeem Hossain, Md. Abdur Rahim. All
authors reviewed the results and approved the final version of the manuscript.

Availability of Data and Materials: The data supporting this study’s findings are available on request from the
corresponding author.

Ethics Approval: Not applicable.

Conflicts of Interest: The authors declare no conflicts of interest to report regarding the present study.
Comput Mater Contin. 2025;82(3) 3677

Abbreviations
AI Artificial intelligence
ML Machine learning
ANN Artificial neural network
DL Deep learning
RL Reinforcement learning
ADAS Advanced driver assistance systems
IoT Internet of Things
PdM Predictive maintenance
NLP Natural language processing
AV Autonomous vehicle
ANI Artificial narrow intelligence
AGI Artificial general intelligence
ASI Artificial super intelligence
SVM Support vector machine
KNN K-nearest neighbour
DT Decision tree
RF Random forest
CNN Convolutional neural network
RNN Recurrent neural network
DQN Deep Q-network
GRU Gated recurrent unit
SI Swarm intelligence
GA Genetic algorithm
ACO Ant colony optimisation
ACC Adaptive cruise control
EV Electric vehicle
UAV Unmanned aerial vehicle
CAN Controller area network
DRL Deep reinforcement learning
GIS Geographical information system
VHMS Vehicle health monitoring system
LSTM Long short-term memory
GAN Generative adversarial network
PSO Particle swarm optimisation
RADAR Radio detection and ranging
LiDAR Light detection and ranging
CA Collision avoidance
Cobot Collaborative robot
SCM Supply chain management
RUL Remaining useful life
AR Augmented reality

References
1. Talaat M, Elkholy MH, Alblawi A, Said T. Artificial intelligence applications for microgrids integration and
management of hybrid renewable energy sources. Artif Intell Rev. 2023;56(9):10557–611. doi:10.1007/s10462-023-
10410-w.
2. Yüksel N, Börklü HR, Sezer HK, Canyurt OE. Review of artificial intelligence applications in engineering design
perspective. Eng Appl Artif Intell. 2023;118(4):105697. doi:10.1016/j.engappai.2022.105697.
3678 Comput Mater Contin. 2025;82(3)

3. Soni N, Sharma EK, Singh N, Kapoor A. Impact of artificial intelligence on businesses: from research, innovation,
market deployment to future shifts in business models. arXiv:1905.02092. 2019.
4. Kamran SS, Haleem A, Bahl S, Javaid M, Prakash C, Budhhi D. Artificial intelligence and advanced materials in
automotive industry: potential applications and perspectives. Mater Today: Proc. 2022;62(5):4207–14. doi:10.1016/
j.matpr.2022.04.727.
5. Al Kuwaiti A, Nazer K, Al-Reedy A, Al-Shehri S, Al-Muhanna A, Subbarayalu AV, et al. A review of the role of
artificial intelligence in healthcare. J Pers Med. 2023;13(6):951. doi:10.3390/jpm13060951.
6. Franki V, Majnarić D, Višković A. A comprehensive review of artificial intelligence (AI) companies in the power
sector. Energies. 2023;16(3):1077. doi:10.3390/en16031077.
7. Kang J, Lee HJ, Jeong SH, Lee HS, Oh KJ. Developing a forecasting model for real estate auction prices using
artificial intelligence. Sustainability. 2020;12(7):2899. doi:10.3390/su12072899.
8. Han S. AI, culture industries and entertainment. In: Elliott A, editor. The social science handbook of AI. London,
UK: Routledge; 2021. p. 295–312.
9. Tarallo E, Akabane GK, Shimabukuro CI, Mello J, Amancio D. Machine learning in predicting demand for fast-
moving consumer goods: an exploratory research. IFAC-PapersOnLine. 2019;52(13):737–42. doi:10.1016/j.ifacol.
2019.11.203.
10. Alhumaid K, Naqbi S, Elsori D, Mansoori M. The adoption of artificial intelligence applications in education. Int
J Data Netw Sci. 2023;7(1):457–66. doi:10.5267/j.ijdns.2022.8.013.
11. Weber FD, Schütte R. State-of-the-art and adoption of artificial intelligence in retailing. Digit Policy Regul Gov.
2019;21(3):264–79. doi:10.1108/DPRG-09-2018-0050.
12. Elbasi E, Mostafa N, AlArnaout Z, Zreikat AI, Cina E, Varghese G, et al. Artificial intelligence technology in the
agricultural sector: a systematic literature review. IEEE Access. 2022 Dec;11(6):171–202. doi:10.1109/ACCESS.2022.
3232485.
13. Alhayani B, Mohammed HJ, Chaloob IZ, Ahmed JS. Effectiveness of artificial intelligence techniques against cyber
security risks apply of IT industry. Mater Today: Proc. 2021. doi:10.1016/j.matpr.2021.02.531.
14. Ukpong EG, Udoh II, Essien IT. Artificial intelligence: opportunities, issues and applications in banking,
accounting, and auditing in Nigeria. Asian J Econ, Bus Account. 2019;8(1):309. doi:10.9734/ajeba/2019/v10i130099.
15. Kortelainen J, Zeb A, Ailisto H. Artificial intelligence for the support of regulator decision making. Espoo, Finland:
VTT Technical Research Centre of Finland; 2020. 30 p.
16. Shi Q, Zhang H. Fault diagnosis of an autonomous vehicle with an improved SVM algorithm subject to unbalanced
datasets. IEEE Trans Ind Electron. 2021;68(7):6248–56. doi:10.1109/TIE.2020.2994868.
17. Luckow A, Kennedy K, Ziolkowski M, Djerekarov E, Cook M, Duffy E, et al. Artificial intelligence and deep
learning applications for automotive manufacturing. In: 2018 IEEE International Conference on Big Data (Big
Data); 2018; Seattle, WA, USA. p. 3144–52.
18. Merabet GH, Essaaidi M, Haddou MB, Qolomany B, Qadir J, Anan M, et al. Intelligent building control systems
for thermal comfort and energy-efficiency: a systematic review of artificial intelligence-assisted techniques. Renew
Sustain Energ Rev. 2021;144:110969. doi:10.1016/j.rser.2021.110969.
19. Dumitrascu O, Dumitrascu M, Dobrota D. Performance evaluation for a sustainable supply chain management
system in the automotive industry using artificial intelligence. Processes. 2020;8(11):1384. doi:10.3390/pr8111384.
20. Lee S, Kim Y, Kahng H, Lee SK, Chung S, Cheong T, et al. Intelligent traffic control for autonomous vehicle systems
based on machine learning. Expert Syst Appl. 2020;144(1):113074. doi:10.1016/j.eswa.2019.113074.
21. Rizvi AT, Haleem A, Bahl S, Javaid M. Artificial intelligence (AI) and its applications in Indian manufacturing: a
review. In: Acharya SK, Mishra DP, editors. Current advances in mechanical engineering: select proceedings of
ICRAMERD 2020; Singapore: Springer; 2021. p. 825–35.
22. Automotive artificial intelligence (AI) market size, share, and trends 2024 to 2034 [Internet]. 2024 [cited 2025 Jan
10]. Available from: https://www.precedenceresearch.com/automotive-artificial-intelligence-market.
23. Li J, Cheng H, Guo H, Qiu S. Survey on artificial intelligence for vehicles. Automot Innov. 2018;1(1):2–14. doi:10.
1007/s42154-018-0009-9.
Comput Mater Contin. 2025;82(3) 3679

24. Bergès C, Céneray C, Albermann G, Bird J, Balakrishnan B, Mustafa SM, et al. Detection and prevention of
assembly defects, by machine learning algorithms, in semiconductor industry for automotive. In: 2020 IEEE
International Symposium on the Physical and Failure Analysis of Integrated Circuits (IPFA); 2020; Singapore. p.
1–9.
25. Fedullo T, Morato A, Tramarin F, Cattini S, Rovati L. Artificial intelligence-based measurement systems for
automotive: a comprehensive review. In: 2022 IEEE International Workshop on Metrology for Automotive
(MetroAutomotive); 2022; Modena, Italy: IEEE. p. 122–7.
26. Rana K, Khatri N. Automotive intelligence: unleashing the potential of AI beyond advance driver assisting system,
a comprehensive review. Comput Electr Eng. 2024;117:109237. doi:10.1016/j.compeleceng.2024.109237.
27. Ammal SM, Kathiresh M, Neelaveni R. Artificial intelligence and sensor technology in the automotive industry:
an overview. In: Kathiresh M, Neelaveni R, editors. Automotive embedded systems: key technologies, innovations,
and applications. Cham: Springer; 2021. p. 145–64.
28. Vermesan O, John R, Pype P, Daalderop G, Kriegel K, Mitic G, et al. Automotive intelligence embedded in
electric connected autonomous and shared vehicles technology for sustainable green mobility. Front Future Transp.
2021;2:688482. doi:10.3389/ffutr.2021.688482.
29. Haenlein M, Kaplan A. A brief history of artificial intelligence: on the past, present, and future of artificial
intelligence. Calif Manage Rev. 2019;61(4):5–14. doi:10.1177/0008125619864925.
30. Bogue R. Robotic vision boosts automotive industry quality and productivity. Ind Robot: Int J. 2013;40(5):415–9.
doi:10.1108/IR-04-2013-342.
31. Tsugawa S, Yatabe T, Hirose T, Matsumoto S. An automobile with artificial intelligence. In: Proceedings of the 6th
International Joint Conference on Artificial Intelligence; 1979; Tokyo, Japan. Vol. 2, p. 893–5.
32. Gage DW. UGV history 101: a brief history of Unmanned Ground Vehicle (UGV) development efforts. In: Naval
command control and ocean surveillance center RDT and E. USA: RDT&E Division, Naval Command, Control
and Ocean Surveillance Center; 1995.
33. Klausmeier R. Using artificial intelligence in vehicle diagnostic systems. In: West Coast International Meeting and
Exposition; 1986 Aug. 12 p.
34. Dougherty M. A review of neural networks applied to transport. Transp Res Part C Emerg Technol.
1995;3(4):247–60. doi:10.1016/0968-090X(95)00009-8.
35. Luu R. A fuzzy logic-based automobile fuel injection control system [M.S. thesis]. Long Beach: Department of
Electrical Engineering, California State University; 1995.
36. Park SJ, Hong S, Kim D, Hussain I, Seo Y. Intelligent in-car health monitoring system for elderly drivers in
connected car. Adv Intell Syst Comput. 2019;823:40–4. doi:10.1007/978-3-319-96074-6.
37. Nikitas A, Michalakopoulou K, Njoya ET, Karampatzakis D. Artificial intelligence, transport and the smart city:
definitions and dimensions of a new mobility era. Sustainability. 2020;12(7):2789. doi:10.3390/su12072789.
38. Abduljabbar R, Dia H, Liyanage S, Bagloee SA. Applications of artificial intelligence in transport: an overview.
Sustainability. 2019;11(1):1–24. doi:10.3390/su11010189.
39. Khayyam H, Javadi B, Jalili M, Jazar RN. Artificial intelligence and internet of things for autonomous vehicles. In:
Jazar R, Dai L, editors. Nonlinear approaches in engineering applications: automotive applications of engineering
problems. Cham: Springer; 2020. p. 39–68.
40. Škavić F. The implementation of artificial intelligence and its future potential. Croatia: Department of Informatics,
University of Zagreb; 2019.
41. Kirwan C, Zhiyong F. Smart city functions. In: Smart cities and artificial intelligence; Amsterdam, Netherlands:
Elsevier; 2020. p. 163–92.
42. Hassani H, Silva ES, Unger S, TajMazinani M, Mac Feely S. Artificial intelligence (AI) or intelligence augmentation
(IA): what is the future? AI. 2020;1(2):143–55. doi:10.3390/ai1020008.
43. Tobaruela JA, Rodríguez AO. Reactive navigation in extremely dense and highly intricate environments. PLoS
One. 2017;12(12):e0189008. doi:10.1371/journal.pone.0189008.
44. Roberts JM, Duff ES, Corke PI. Reactive navigation and opportunistic localization for autonomous underground
mining vehicles. Inf Sci. 2002;145(1–2):127–46. doi:10.1016/S0020-0255(02)00227-X.
3680 Comput Mater Contin. 2025;82(3)

45. Gupta A, Anpalagan A, Guan L, Khwaja AS. Deep learning for object detection and scene perception in self-driving
cars: survey, challenges, and open issues. Array. 2021;10(10):100057. doi:10.1016/j.array.2021.100057.
46. Andreu-Perez J, Deligianni F, Ravi D, Yang GZ. Artificial intelligence and robotics. arXiv:1803.10813. 2018.
47. Surden H, Williams MA. Technological opacity, predictability, and self-driving cars. Cardozo L Rev. 2016;38:121.
doi:10.2139/ssrn.2747491.
48. Rabinowitz NC, Perbet F, Song HF, Zhang C, Eslami SMA, Botvinick M. Machine theory of mind. In: 35th
International Conference on Machine Learning; 2018; Stockholm, Sweden. p. 4218–27.
49. Schlatow J, Moostl M, Ernst R, Nolte M, Jatzkowski I, Maurer M. Self-awareness in autonomous automotive
systems. In: Design, Automation & Test in Europe Conference & Exhibition (DATE); 2017; Lausanne, Switzerland:
IEEE. p. 1050–5. doi:10.23919/DATE.2017.7927145.
50. Shi D, Zhao J, Eze C, Wang Z, Wang J, Lian Y, et al. Cloud-based artificial intelligence framework for battery
management system. Energies. 2023;16(11):4403. doi:10.3390/en16114403.
51. Nori R, Zucchelli MM, Cordellieri P, Quaglieri A, Palmiero M, Guariglia P, et al. The prevention of road accidents
in non-expert drivers: exploring the influence of theory of mind and driving style. Saf Sci. 2024;175(11):106516.
doi:10.1016/j.ssci.2024.106516.
52. Shao W, Li J, Wang H. Self-aware trajectory prediction for safe autonomous driving. In: 2023 IEEE Intelligent
Vehicles Symposium (IV); 2023; Anchorage, AK, USA: IEEE; p. 1–8.
53. Gryth E, Rundberg D. Am AI ready? Investigating the impacts of artificial intelligence on [M.S. thesis]. Sweden:
School of Business, Economics, and Law, University of Gothenburg; 2018.
54. Li L, Zheng N, Wang FY. A theoretical foundation of intelligence testing and its application for intelligent vehicles.
IEEE Trans Intell Transp Syst. 2020;22(10):6297–306. doi:10.1109/TITS.2020.2991039.
55. Chai TY, Nizam I. Impact of artificial intelligence in automotive industries transformation. Int J Inf Syst Eng.
2021;9(1):1–35. doi:10.24924/ijise/2021.04/v9.iss2/01.35.
56. Rahim MA, Rahman MM, Rahman MA, Muzahid AJM, Kamarulzaman SF. A framework of IoT-enabled vehicular
noise intensity monitoring system for smart city. In: Advances in robotics, automation and data analytics: selected
papers from ICITES 2020. Switzerland: Springer; 2021. p. 194–205.
57. Ali S, Adeel M, Johar S, Zeeshan M, Baseer S, Irshad A, Arif M. Machine learning technologies for secure vehicular
communication in internet of vehicles: recent advances and applications. Secur Commun Netw. 2021;2021(1):1–23.
doi:10.1155/2021/9609823.
58. Harold CKD, Prakash S, Hofman T. Powertrain control for hybrid-electric vehicles using supervised machine
learning. Vehicles. 2020;2(2):267–86. doi:10.3390/vehicles2020015.
59. Chouhad H, El Mansori M, Knoblauch R, Corleto C. Smart data driven defect detection method for surface quality
control in manufacturing. Meas Sci Technol. 2021;32(10):105403. doi:10.1088/1361-6501/ac0b6c.
60. Panahandeh G, Ek E, Mohammadiha N. Road friction estimation for connected vehicles using supervised machine
learning. In: IEEE Intelligent Vehicles Symposium; 2017; Los Angeles, CA, USA. p. 1262–7.
61. El-Sayed H, Sankar S, Daraghmi Y-A, Tiwari P, Rattagan E, Mohanty M, et al. Accurate traffic flow prediction in
heterogeneous vehicular networks in an intelligent transport system using a supervised non-parametric classifier.
Sensors. 2018;18(6):1696. doi:10.3390/s18061696.
62. Cuenca LG, Sanchez-Soriano J, Puertas E, Aliane N, Andrés JF. Machine learning techniques for undertaking
roundabouts in autonomous driving. Sensors. 2019;19(10):2386. doi:10.3390/s19102386.
63. Yousefi P, Fekriazgomi H, Demir MA, Prevost JJ, Jamshidi M. Data-driven fault detection of un-manned aerial
vehicles using supervised learning over cloud networks. In: 2018 World Automation Congress (WAC); 2018;
Stevenson, WA, USA: IEEE. p. 1–6.
64. Lourens M, Sharma S, Pulugu R, Gehlot A, Manoharan G, Kapila D. Machine learning-based predictive analytics
and big data in the automotive sector. In: 2023 3rd International Conference on Advance Computing and
Innovative Technologies in Engineering (ICACITE); 2023; Greater Noida, India: IEEE. p. 1043–8.
65. Hegde J, Rokseth B. Applications of machine learning methods for engineering risk assessment—a review. Saf Sci.
2020;122:104492. doi:10.1016/j.ssci.2019.09.015.
Comput Mater Contin. 2025;82(3) 3681

66. Madasamy J, Parameswari R. Data mining and machine learning techniques (mahout on hadoop) for automotive
industries: a survey. Int J Pure Appl Math. 2018;119(10):1–5.
67. Naeem S, Ali A, Anam S, Ahmed MM. An unsupervised machine learning algorithms: comprehensive review. Int
J Comput Digit Syst. 2023;13(1):911–21. doi:10.12785/ijcds/130172.
68. Usmani UA, Happonen A, Watada J. A review of unsupervised machine learning frameworks for anomaly
detection in industrial applications. In: Intelligent computing; Cham: Springer; 2022. p. 158–89.
69. Garay JM, DIedrich C. Analysis of the applicability of fault detection and failure prediction based on unsuper-
vised learning and monte carlo simulations for real devices in the industrial automobile production. In: IEEE
International Conference on Industrial Informatics (INDIN); 2019; Helsinki, Finland. p. 1279–84.
70. Shaeiri Z, Kazemitabar SJ. Fast unsupervised automobile insurance fraud detection based on spectral ranking of
anomalies. Int J Eng. 2020;33(7):1–5.
71. Ezeobi U, Olufowobi H, Young C, Zambreno J, Bloom G. Reverse engineering controller area network messages
using unsupervised machine learning. IEEE Consum Electron Mag. 2020;11(1):50–6. doi:10.1109/MCE.2020.
3023538.
72. Watanabe H, Maly T, Wallner J, Prokop G. Cluster-linkage analysis in traffic data clustering for development
of advanced driver assistance systems. In: 2020 3rd International Conference on Information and Computer
Technologies (ICICT); 2020; San Jose, CA, USA. p. 54–9.
73. Dupre R, Fajtl J, Argyriou V, Remagnino P. Improving dataset volumes and model accuracy with semi-supervised
iterative self-learning. IEEE Trans Image Process. 2019;29:4337–48. doi:10.1109/TIP.2019.2913986.
74. Lamm N, Jaiprakash S, Srikanth M, Drori I. Vehicle trajectory prediction by transfer learning of semi-supervised
models. arXiv:2007.06781. 2020.
75. Shibly KH, Hossain MD, Inoue H, Taenaka Y, Kadobayashi Y. A feature-aware semi-supervised learning approach
for automotive ethernet. In: 2023 IEEE International Conference on Cyber Security and Resilience (CSR); 2023;
Venice, Italy: IEEE. p. 426–31.
76. Hoang TN, Kim D. Detecting in-vehicle intrusion via semi-supervised learning-based convolutional adversarial
autoencoders. Veh Commun. 2022;38(2):100520. doi:10.1016/j.vehcom.2022.100520.
77. Liu Z, Huang S, Jin W, Mu Y. Broad learning system for semi-supervised learning. Neurocomputing.
2021;444(7553):38–47. doi:10.1016/j.neucom.2021.02.059.
78. Kiran BR, Sobh I, Talpaert V, Mannion P, Al Sallab AA, Yogamani S, et al. Deep reinforcement learning for
autonomous driving: a survey. IEEE Trans Intell Transp Syst. 2021;23(6):4909–26. doi:10.1109/TITS.2021.3054625.
79. Navarro A, Genc S, Rangarajan P, Khalil R, Goberville N, Rojas JF, et al. Using reinforcement learning and
simulation to develop autonomous vehicle control strategies. SAE Technical Papers. 2020;1:1–8. doi:10.4271/2020-
01-0737.
80. Nazari M, Oroojlooy A, Takáč M, Snyder LV. Reinforcement learning for solving the vehicle routing problem. In:
32nd Conference on Neural Information Processing Systems (NeurIPS 2018); 2018; Montréal, QC, Canada. p. 1–11.
81. Ye Y, Zhang X, Sun J. Automated vehicle’s behavior decision making using deep reinforcement learning and
high-fidelity simulation environment. Transp Res Part C Emerg Technol. 2019;107(6):155–70. doi:10.1016/j.trc.2019.
08.011.
82. Zeng Y, Xu X, Jin S, Zhang R. Simultaneous navigation and radio mapping for cellular-connected UAV with deep
reinforcement learning. IEEE Trans Wirel Commun. 2021;20(7):4205–20. doi:10.1109/TWC.2021.3056573.
83. Huang H, Yang Y, Wang H, Ding Z, Sari H, Adachi F. Deep reinforcement learning for UAV navigation through
massive MIMO technique. IEEE Trans Veh Technol. 2020;69(1):1117–21. doi:10.1109/TVT.2019.2952549.
84. Polvara R, Sharma S, Wan J, Manning A, Sutton R. Autonomous vehicular landings on the deck of an unmanned
surface vehicle using deep reinforcement learning. Robotica. 2019;37(11):1867–82. doi:10.1017/S0263574719000316.
85. Koch L, Roeser D, Badalian K, Lieb A, Andert J. Cloud-based reinforcement learning in automotive control
function development. Vehicles. 2023;5(3):914–30. doi:10.3390/vehicles5030050.
86. Yan Z, Kreidieh AR, Vinitsky E, Bayen AM, Wu C. Unified automatic control of vehicular systems with
reinforcement learning. IEEE Trans Autom Sci Eng. 2022;20(2):789–804. doi:10.1109/TASE.2022.3168621.
3682 Comput Mater Contin. 2025;82(3)

87. Xu X, Zuo L, Li X, Qian L, Ren J, Sun Z. A reinforcement learning approach to autonomous decision making of
intelligent vehicles on highways. IEEE Trans Syst Man Cybern: Syst. 2018;50(10):3884–97.
88. Schoch A, Refflinghaus R, Zivkovic P. Deep learning based predictive testing strategy in the automotive industry.
Procedia CIRP. 2023;118(15–16):1108–13. doi:10.1016/j.procir.2023.06.190.
89. Chen C, Liu Y, Sun X, Cairano-Gilfedder CD, Titmus S. Automobile maintenance prediction using deep learning
with gis data. Procedia CIRP. 2019;81:447–52. doi:10.1016/j.procir.2019.03.077.
90. Naqvi RA, Arsalan M, Batchuluun G, Yoon HS, Park KR. Deep learning-based gaze detection system for
automobile drivers using a NIR camera sensor. Sensors. 2018;18(2):1–34. doi:10.3390/s18020456.
91. Ulrich M, Gläser C, Timm F. Deepreflecs: deep learning for automotive object classification with radar reflections.
In: 2021 IEEE Radar Conference (RadarConf21); 2021; Atlanta, GA, USA: IEEE. p. 1–6.
92. Paparusso L, Melzi S, Braghin F. A deep-learning framework to predict the dynamics of a human-driven vehicle
based on the road geometry. arXiv:2103.03825v1. 2021.
93. Tammewar A, Chaudhari N, Saini B, Venkatesh D, Dharahas G, Vora D, et al. Improving the performance
of autonomous driving through deep reinforcement learning. Sustainability. 2023;15(18):13799. doi:10.3390/
su151813799.
94. Talpaert V, Sobh I, Kiran BR, Mannion P, Yogamani S, El-Sallab A, et al. Exploring applications of deep
reinforcement learning for real-world autonomous driving systems. arXiv:1901.01536. 2019.
95. Ren J, Gaber H, Al Jabar SS. Applying deep learning to autonomous vehicles: a survey. In: 2021 4th International
Conference on Artificial Intelligence and Big Data (ICAIBD); 2021; Chengdu, China: IEEE. p. 247–52.
96. Singh KB, Arat MA. Deep learning in the automotive industry: recent advances and application examples.
arXiv:1906.08834. 2019.
97. Wang J, Ma Y, Zhang L, Gao RX, Wu D. Deep learning for smart manufacturing: methods and applications. J
Manuf Syst. 2018;48(2):144–56. doi:10.1016/j.jmsy.2018.01.003.
98. Chandriah KK, Naraganahalli RV. RNN/LSTM with modified Adam optimizer in deep learning approach for
automobile spare parts demand forecasting. Multimed Tools Appl. 2021;80(17):26145–59. doi:10.1007/s11042-021-
10913-0.
99. Osco LP, Junior JM, Ramos APM, de Castro Jorge LA, Fatholahi SN, de Andrade Silva J, et al. A review on deep
learning in UAV remote sensing. Int J Appl Earth Obs Geoinf. 2021;102:102456. doi:10.1016/j.jag.2021.102456.
100. Lindfield G, Penny J. Optimization methods. In: Numerical methods. Amsterdam, The Netherlands: Elsevier; 2019.
p. 433–83.
101. Santini A, Schneider M, Vidal T, Vigo D. Decomposition strategies for vehicle routing heuristics. INFORMS J
Comput. 2023;35(3):543–59. doi:10.1287/ijoc.2023.1288.
102. Homayouni Z, Pishvaee MS, Jahani H, Ivanov D. A robust-heuristic optimization approach to a green supply
chain design with consideration of assorted vehicle types and carbon policies under uncertainty. Ann Oper Res.
2023;324(1–2):395–435. doi:10.1007/s10479-021-03985-6.
103. Pérez MAJ, Loaiza REP, Flores PMQ, Ponce OA, Peralta CF. A heuristic algorithm for the routing and scheduling
problem with time windows: a case study of the automotive industry in Mexico. Algorithms. 2019;12(5):1–19. doi:10.
3390/a12050111.
104. Gunantara N, Putra IDNN. The characteristics of metaheuristic method in selection of path pairs on multicriteria
ad hoc networks. J Comput Netw Commun. 2019;2019:1–6. doi:10.1155/2019/7983583.
105. Rajwar K, Deep K, Das S. An exhaustive review of the metaheuristic algorithms for search and optimization:
Taxonomy, applications, and open challenges. Artif Intell Rev. 2023;56(11):13187–257. doi:10.1007/s10462-023-
10470-y.
106. Sharma M, Kaur P. A comprehensive analysis of nature-inspired meta-heuristic techniques for feature selection
problem. Arch Comput Methods Eng. 2021;28(3):1103–27. doi:10.1007/s11831-020-09412-6.
107. Tang J, Liu G, Pan Q. A review on representative swarm intelligence algorithms for solving optimization problems:
applications and trends. IEEE/CAA J Automatica Sinica. 2021;8(10):1627–43. doi:10.1109/JAS.2021.1004129.
Comput Mater Contin. 2025;82(3) 3683

108. Shi K, Huang L, Jiang D, Sun Y, Tong X, Xie Y, et al. Path planning optimization of intelligent vehicle based on
improved genetic and ant colony hybrid algorithm. Front Bioeng Biotechnol. 2022;10:905983. doi:10.3389/fbioe.
2022.905983.
109. Kyriakakis NA, Marinaki M, Marinakis Y. A hybrid ant colony optimization-variable neighborhood descent
approach for the cumulative capacitated vehicle routing problem. Comput Operat Res. 2021;134:105397. doi:10.
1016/j.cor.2021.105397.
110. Hashim A, Saini T, Bhardwaj H, Jothi A, Kumar AV. Application of swarm intelligence in autonomous cars
for obstacle avoidance. In: Krishna A, Srikantaiah K, Naveena C, editors. Studies in computational intelligence.
Singapore: Springer; 2019.
111. Jiang C, Fu J, Liu W. Research on vehicle routing planning based on adaptive ant colony and particle swarm
optimization algorithm. Int J Intell Transp Syst Res. 2021;19(1):83–91. doi:10.1007/s13177-020-00224-3.
112. Xiang X, Qiu J, Xiao J, Zhang X. Demand coverage diversity based ant colony optimization for dynamic vehicle
routing problems. Eng Appl Artif Intell. 2020;91(5):103582. doi:10.1016/j.engappai.2020.103582.
113. Kaushal P, Khurana M, Ramkumar KR. A systematic review of swarm intelligence algorithms to perform routing
for VANETs communication. ECS Trans. 2022;107(1):5027–35. doi:10.1149/10701.5027ecst.
114. Faisal M, Albogamy F. Ant colony optimization algorithm enhancement for better performance. In: 2023 IEEE
world AI IoT congress (AIIoT); 2023; Seattle, WA, USA: IEEE. p. 701–10.
115. Eboli MG, Maberry CM, Gibbs IA, Haddad R. Detecting potential vehicle concerns using natural language
processing applied to automotive big data. In: 26th International Technical Conference on the Enhanced Safety of
Vehicles (ESV): Technology: Enabling a Safer Tomorrow; 2019; Eindhoven, Netherlands.
116. Masello L, Castignani G, Sheehan B, Murphy F, McDonnell K. On the road safety benefits of advanced driver
assistance systems in different driving contexts. Transp Res Interdiscip Perspect. 2022;15:100670. doi:10.1016/j.trip.
2022.100670.
117. Tateno S, Zhu Y, Meng F. Hand gesture recognition system for in-car device control based on infrared array
sensor. In: 2019 58th Annual Conference of the Society of Instrument and Control Engineers of Japan (SICE); 2019;
Hiroshima, Japan: IEEE. p. 701–6.
118. Jaradat GA, Alzubaidi MA, Otoom M. A novel human-vehicle interaction assistive device for Arab drivers using
speech recognition. IEEE Access. 2022;10:127514–29. doi:10.1109/ACCESS.2022.3226539.
119. Eksin C, Swenson B, Kar S, Ribeiro A. Game theoretic learning. In: Cooperative and graph signal processing.
Amsterdam, The Netherlands: Elsevier; 2018. p. 209–35.
120. Yuan M, Shan J, Mi K. Deep reinforcement learning based game-theoretic decision-making for autonomous
vehicles. IEEE Robot Autom Lett. 2021;7(2):818–25. doi:10.1109/LRA.2021.3134249.
121. Lopez VG, Lewis FL, Liu M, Wan Y, Nageshrao S, Filev D. Game-theoretic lane-changing decision making and
payoff learning for autonomous vehicles. IEEE Trans Veh Technol. 2022;71(4):3609–20. doi:10.1109/TVT.2022.
3148972.
122. Wang M, Wang Z, Talbot J, Gerdes JC, Schwager M. Game-theoretic planning for self-driving cars in multivehicle
competitive scenarios. IEEE Trans Robot. 2021;37(4):1313–25. doi:10.1109/TRO.2020.3047521.
123. Chandra R, Manocha D. Game-theoretic multi-agent planning with human drivers at intersections, roundabouts,
and merging. IEEE Robot Autom Lett. 2022;7(2):2676–83. doi:10.1109/LRA.2022.3144516.
124. Ardekani AA, Chahe A, Yazdi MRH. Combining deep learning and game theory for path planning in autonomous
racing cars. In: 2022 10th RSI International Conference on Robotics and Mechatronics (ICRoM); 2022; Tehran,
Iran: IEEE. p. 564–71.
125. Hossain MN, Rahman MM, Ramasamy D, Kadirgama K, Noor MM. Advancements, challenges, and implications
for navigating the autonomous vehicle revolution. J Mech Eng Sci. 2024;8(2):10077–93. doi:10.15282/jmes.18.2.2024.
9.0796.
126. Salais-Fierro TE, Saucedo-Martinez JA, Rodriguez-Aguilar R, Vela-Haro JM. Demand prediction using a soft-
computing approach: a case study of automotive industry. Appl Sci. 2020;10(3):829. doi:10.3390/app10030829.
127. Castro DM, Silv Parreiras F. A review on multi-criteria decision-making for energy efficiency in automotive
engineering. Appl Comput Inform. 2021;17(1):53–78. doi:10.1016/j.aci.2018.04.004.
3684 Comput Mater Contin. 2025;82(3)

128. Godina R, Silva BGR, Espadinha-Cruz P. A DMAIC integrated fuzzy FMEA model: a case study in the Automotive
Industry. Appl Sci. 2021;11(8):3726. doi:10.3390/app11083726.
129. Çeven S, Albayrak A, Bayır R. Real-time range estimation in electric vehicles using fuzzy logic classifier. Comput
Electr Eng. 2020;83(0):106577. doi:10.1016/j.compeleceng.2020.106577.
130. Wankhede VA, Vinodh S. Benchmarking Industry 4.0 readiness evaluation using fuzzy approaches. Benchmark-
ing: Int J. 2023;30(1):281–306. doi:10.1108/BIJ-08-2021-0505.
131. Mateichyk V, Kostian N, Smieszek M, Mosciszewski J, Tarandushka L. Evaluating vehicle energy efficiency in urban
transport systems based on fuzzy logic models. Energies. 2023;16(2):734. doi:10.3390/en16020734.
132. Sathyan R, Parthiban P, Dhanalakshmi R, Minz A. A combined big data analytics and Fuzzy DEMATEL
technique to improve the responsiveness of automotive supply chains. J Ambient Intell Humaniz Comput.
2021;12(7):7949–63. doi:10.1007/s12652-020-02524-8.
133. Ning H, Yin R, Ullah A, Shi F. A survey on hybrid human-artificial intelligence for autonomous driving. IEEE
Trans Intell Transp Syst. 2021;23(7):6011–26. doi:10.1109/TITS.2021.3074695.
134. Al Shukairi H, Cardoso RC. ML-MAS: a hybrid AI framework for self-driving vehicles. In: AAMAS’23: Proceed-
ings of the 2023 International Conference on Autonomous Agents and Multiagent Systems; 2023; London, UK:
International Foundation for Autonomous Agents and Multiagent Systems (IFAAMAS). p. 1191–9.
135. Rahim MA, Rahman MM, Islam MS, Muzahid AJM, Rahman MA, Ramasamy D. Deep learning-based vehicular
engine health monitoring system utilising a hybrid convolutional neural network/bidirectional gated recurrent
unit. Expert Syst Appl. 2024;257(3):125080. doi:10.1016/j.eswa.2024.125080.
136. Al-Saadi Z, Phan Van D, Moradi Amani A, Fayyazi M, Sadat Sajjadi S, Ba Pham D, et al. Intelligent driver assistance
and energy management systems of hybrid electric autonomous vehicles. Sustainability. 2022;14(15):9378. doi:10.
3390/su14159378.
137. Nazari S, Gowans N, Abtahi M. Powertrain hybridization for autonomous vehicles. arXiv:2309.17442. 2023.
138. Al-Jumaili MH, Özok YE, Ibrahim AA, Bayat O. A robust hybrid control model implementation for autonomous
vehicles. In: 2024 8th International Artificial Intelligence and Data Processing Symposium (IDAP); 2024; Malatya,
Turkiye: IEEE. p. 1–5.
139. Reda A, Benotsmane R, Bouzid A, Vásárhelyi J. A hybrid machine learning-based control strategy for autonomous
driving optimization. Acta Polytechnica Hung. 2023;20(9):165–86. doi:10.12700/APH.20.9.2023.9.10.
140. Phan D, Bab-Hadiashar A, Fayyazi M, Hoseinnezhad R, Jazar RN, Khayyam H. Interval type 2 fuzzy logic control
for energy management of hybrid electric autonomous vehicles. IEEE Trans Intell Veh. 2020;6(2):210–20. doi:10.
1109/TIV.2020.3011954.
141. Santana EFZ, Covas G, Duarte F, Santi P, Ratti C, Kon F. Transitioning to a driverless city: evaluating a hybrid
system for autonomous and non-autonomous vehicles. Simul Model Pract Theory. 2021;107(28):102210. doi:10.1016/
j.simpat.2020.102210.
142. Marina LA, Trasnea B, Grigorescu SM. A multi-platform framework for artificial intelligence engines in automotive
systems. In: 2018 22nd International Conference on System Theory, Control and Computing (ICSTCC); 2018;
Sinaia, Romania: IEEE. p. 559–64.
143. Ruggaber J, Ahmic K, Brembeck J, Baumgartner D, Tobolář J. AI-for-mobility—a new research platform for AI-
based control methods. Appl Sci. 2023;13(5):2879. doi:10.3390/app13052879.
144. Díaz AA, Kurka PRG. Computer vision methods for automotive applications. Tecnia. 2020;30(2):74–81. doi:10.
21754/tecnia.v30i2.801.
145. Nagy M, Lăzăroiu G. Computer vision algorithms, remote sensing data fusion techniques, and mapping and
navigation tools in the Industry 4.0-based Slovak automotive sector. Mathematics. 2022;10(19):3543. doi:10.3390/
math10193543.
146. Nkuzo L, Sibiya M, Markus E. Computer vision-based applications in modern cars for safety purposes: a systematic
literature review. In: 2023 Conference on Information Communications Technology and Society (ICTAS); 2023;
Durban, South Africa: IEEE. p. 1–9.
Comput Mater Contin. 2025;82(3) 3685

147. Konstantinidis FK, Mouroutsos SG, Gasteratos A. The role of machine vision in industry 4.0: an automotive
manufacturing perspective. In: 2021 IEEE International Conference on Imaging Systems and Techniques (IST);
2021; Kaohsiung, Taiwan: IEEE. p. 1–6.
148. Abagiu MM, Cojocaru D, Manta FL, Mariniuc A. Detection of a surface defect on an engine block using computer
vision. In: 2021 22nd International Carpathian Control Conference (ICCC); 2021; Velké Karlovice, Czech Republic:
IEEE. p. 1–5.
149. Feng MQ, Leung RY. Application of computer vision for estimation of moving vehicle weight. IEEE Sens J.
2020;21(10):11588–97. doi:10.1109/JSEN.2020.3038186.
150. Darapaneni N, Ravikumar M, Singh S, Tiwari A, Das S, Paduri AR, et al. Computer vision application in automobile
error detection. In: 2022 Interdisciplinary Research in Technology and Management (IRTM); 2022; Kolkata, India:
IEEE. p. 1–7.
151. Mahadevkar SV, Khemani B, Patil S, Kotecha K, Vora DR, Abraham A, et al. A review on machine learning styles
in computer vision—techniques and future directions. IEEE Access. 2022;10(1):107293–329. doi:10.1109/ACCESS.
2022.3209825.
152. Srivastava N, Lamba T, Agarwal M. Comparative analysis of different machine learning techniques. In: Futuristic
Trends in Networks and Computing Technologies (FTNCT 2019); 2020; Chandigarh, India: Springer. p. 245–55.
153. Voulodimos A, Doulamis N, Doulamis A, Protopapadakis E. Deep learning for computer vision: a brief review.
Comput Intell Neurosci. 2018;2018(1):7068349. doi:10.1155/2018/7068349.
154. Ma W, Cheng F, Liu Y. Deep-learning-enabled on-demand design of chiral metamaterials. ACS Nano.
2018;12(6):6326–34. doi:10.1021/acsnano.8b03569.
155. Mao Y, Miller RA, Bodenreider O, Nguyen V, Fung KW. Two complementary AI approaches for predicting UMLS
semantic group assignment: heuristic reasoning and deep learning. J Am Med Inform Assoc. 2023;30(12):1887–94.
doi:10.1093/jamia/ocad152.
156. ul Hassan CA, Khan MS, Irfan R, Iqbal J, Hussain S, Sajid Ullah S, et al. Optimizing deep learning model
for software cost estimation using hybrid meta-heuristic algorithmic approach. Comput Intell Neurosci.
2022;2022(1):3145956. doi:10.1155/2022/3145956.
157. Karimi-Mamaghan M, Mohammadi M, Meyer P, Karimi-Mamaghan AM, Talbi EG. Machine learning at the
service of meta-heuristics for solving combinatorial optimization problems: a state-of-the-art. Eur J Oper Res.
2022;296(2):393–422. doi:10.1016/j.ejor.2021.04.032.
158. Chiroma H, Gital AY, Rana N, Abdulhamid SM, Muhammad AN, Umar AY, et al. Nature inspired meta-heuristic
algorithms for deep learning: recent progress and novel perspective. In: Advances in Computer Vision: Proceedings
of the 2019 Computer Vision Conference (CVC); 2020; Las Vegas, NV, USA. Cham: Springer. Vol. 11, p. 59–70.
159. Dellermann D, Ebel P, Söllner M, Leimeister JM. Hybrid intelligence. Bus Inf Syst Eng. 2019;61(5):637–43. doi:10.
1007/s12599-019-00595-2.
160. Sun Z, Liu Y, Wang J, Li G, Anil C, Li K, et al. Applications of game theory in vehicular networks: a survey. IEEE
Commun Surv Tutor. 2021;23(4):2660–710. doi:10.1109/COMST.2021.3108466.
161. Lü X, Long L, Deng R, Meng R. Image feature extraction based on fuzzy restricted Boltzmann machine.
Measurement. 2022;204(2):112063. doi:10.1016/j.measurement.2022.112063.
162. Soltanali H, Rohani A, Tabasizadeh M, Abbaspour-Fard MH, Parida A. An improved fuzzy inference system-based
risk analysis approach with application to automotive production line. Neural Comput Appl. 2020;32(14):10573–91.
doi:10.1007/s00521-019-04593-z.
163. Varghese NV, Mahmoud QH. A hybrid multi-task learning approach for optimizing deep reinforcement learning
agents. IEEE Access. 2021;9:44681–703. doi:10.1109/ACCESS.2021.3065710.
164. Phatai G, Luangrungruang T. A comparative study of hybrid neural network with metaheuristics for student
performance classification. In: 2023 11th International Conference on Information and Education Technology
(ICIET); 2023; Fujisawa, Japan: IEEE. p. 448–52.
165. Manivarma D, Akilandeswari A, Radhika S. Performance analysis of vehicle detection using support vector
machine comparing with K-nearest neighbor algorithm. In: 2023 International Conference on Advances in
Computing, Communication and Applied Informatics (ACCAI); 2023; Chennai, India: IEEE. p. 1–4.
3686 Comput Mater Contin. 2025;82(3)

166. Singh S, Gill R, Kumar R, Shrivastava A, Srivastava AP, Thakur G. A comparative analysis of machine learning
algorithms for predictive maintenance in electrical systems. In: 2024 4th International Conference on Innovative
Practices in Technology and Management (ICIPTM); 2024; Noida, India: IEEE. p. 1–6.
167. Awan FM, Saleem Y, Minerva R, Crespi N. A comparative analysis of machine/deep learning models for parking
space availability prediction. Sensors. 2020;20(1):322. doi:10.3390/s20010322.
168. Hickling T, Aouf N, Spencer P. Robust adversarial attacks detection based on explainable deep reinforcement
learning for UAV guidance and planning. IEEE Trans Intell Vehicles. 2023;8(10):4381–94. doi:10.1109/TIV.2023.
3296227.
169. Thammachantuek I, Kosolsomnbat S, Ketcham M. Comparison of machine learning algorithm’s performance
based on decision making in autonomous car. In: 2018 International Joint Symposium on Artificial Intelligence
and Natural Language Processing (iSAI-NLP); 2018; Pattaya, Thailand: IEEE. p. 1–6.
170. Liu C, Tang D, Zhu H, Nie Q. A novel predictive maintenance method based on deep adversarial learning in the
intelligent manufacturing system. IEEE Access. 2021;9:49557–75. doi:10.1109/ACCESS.2021.3069256.
171. Dong T, Qi Q, Wang J, Liu AX, Sun H, Zhuang Z, et al. Generative adversarial network-based transfer reinforcement
learning for routing with prior knowledge. IEEE Trans Netw Serv Manag. 2021;18(2):1673–89. doi:10.1109/TNSM.
2021.3077249.
172. Minaeva A, Roy D, Akesson B, Hanzálek Z, Chakraborty S. Control performance optimization for application
integration on automotive architectures. IEEE Trans Comput. 2020;70(7):1059–73. doi:10.1109/TC.2020.3003083.
173. Hu Z, Huang J, Yang Z, Zhong Z. Cooperative-game-theoretic optimal robust path tracking control for autonomous
vehicles. J Vib Control. 2022;28(5–6):520–35. doi:10.1177/10775463211009383.
174. Guo Z, Sun D, Zhou L. Game algorithm of intelligent driving vehicle based on left-turn scene of crossroad traffic
flow. Comput Intell Neurosci. 2022;2022(1):9318475. doi:10.1155/2022/9318475.
175. Tripathi S, Shrivastava A, Jana KC. Chimp optimization-based fuzzy controller for hybrid electric vehicle speed
control using electronic throttle plate. Optim Control Appl Methods. 2024;45(1):163–84. doi:10.1002/oca.3051.
176. Tordecilla RD, do Martins LC, Panadero J, Copado PJ, Perez-Bernabeu E, Juan AA. Fuzzy simheuristics for
optimizing transportation systems: dealing with stochastic and fuzzy uncertainty. Appl Sci. 2021;11(17):7950. doi:10.
3390/app11177950.
177. Takialddin AS. Artificial Intelligent Technology for safe driver assistance system. Int J Comput Aided Eng Technol.
2020;13(1–2):183–91. doi:10.1504/IJCAET.2020.108112.
178. Iyer LS. AI enabled applications towards intelligent transportation. Transp Eng. 2021;5(189):100083. doi:10.1016/j.
treng.2021.100083.
179. Kukkala VK, Tunnell J, Pasricha S, Bradley T. Advanced driver-assistance systems: a path toward autonomous
vehicles. IEEE Consum Electron Mag. 2018;7(5):18–25. doi:10.1109/MCE.2018.2828440.
180. Wu D, Zhu H, Zhu Y, Chang V, He C, Hsu CH, et al. Anomaly detection based on RBM-LSTM neural network for
CPS in advanced driver assistance system. ACM Tran Cyber-Phys Syst. 2020;4(3):1–17. doi:10.1145/3377408.
181. Hasenjäger M, Heckmann M, Wersing H. A survey of personalization for advanced driver assistance systems. IEEE
Trans Intell Veh. 2019;5(2):335–44. doi:10.1109/TIV.2019.2955910.
182. De-Las-Heras G, Sánchez-Soriano J, Puertas E. Advanced driver assistance systems (ADAS) based on
machine learning techniques for the detection and transcription of variable message signs on roads. Sensors.
2021;21(17):5866. doi:10.3390/s21175866.
183. Salve S, Jadhav T, Gajare S, Patil PH. Advanced driver assistance system for drivers using machine learning and
artificial intelligence techniques. Int J Res Appl Sci Eng Technol. 2022;10(5):3885–92. doi:10.22214/ijraset.2022.
43260.
184. Antony MM, Whenish R. Advanced driver assistance systems (ADAS). In: Kathiresh M, Neelaveni R, editors.
Automotive embedded systems: key technologies, innovations, and applications. Cham: Springer; 2021. p. 165–81.
185. Elmorshedy L, Abdulhai B, Kamel I. Quantitative evaluation of the impacts of the time headway of adaptive cruise
control systems on congested urban freeways using different car following models and early control results. IEEE
Open J Intell Transp Syst. 2022;3(4):288–301. doi:10.1109/OJITS.2022.3166394.
Comput Mater Contin. 2025;82(3) 3687

186. Yao S, Guan R, Huang X, Li Z, Sha X, Yue Y, et al. Radar-camera fusion for object detection and semantic
segmentation in autonomous driving: a comprehensive review. IEEE Trans Intell Veh. 2023;9(1):2094–128. doi:10.
1109/TIV.2023.3307157.
187. Mahadika P, Subiantoro A, Kusumoputro B. Neural network predictive control approach design for adaptive cruise
control. Int J Technol. 2020;11(7):1451. doi:10.14716/ijtech.v11i7.4592.
188. Suleiman L, Vlasov S. Research on the impact of different controllers on the adapative cruise control system.
In: 2023 International Russian Automation Conference (RusAutoCon); 2023; Sochi, Russian Federation: IEEE. p.
634–9.
189. Nava D, Panzani G, Zampieri P, Savaresi SM. A personalized adaptive cruise control driving style characterization
based on a learning approach. In: 2019 IEEE Intelligent Transportation Systems Conference (ITSC); 2019;
Auckland, New Zealand: IEEE. p. 2901–6.
190. Zhang L, Liu C, Wang B, Chen X. An innovative method for scene recognition and target tracking for
adaptive cruise control in complex traffic environment. Proc Inst Mech Eng, Part D: J Automobile Eng.
2024;238(7):1896–912. doi:10.1177/09544070231156956.
191. Yin W. Machine learning for adaptive cruise control target selection [dissertation]. 2019. Sweden: School of
Electrical Engineering and Computer Science, KTH.
192. Rill RA, Faragó KB. Collision avoidance using deep learning-based monocular vision. SN Comput Sci.
2021;2(5):375. doi:10.1007/s42979-021-00759-6.
193. Yu J, Petnga L. Space-based collision avoidance framework for autonomous vehicles. Procedia Comput Sci.
2018;140(12):37–45. doi:10.1016/j.procs.2018.10.290.
194. Kotur M, Lukić N, Krunić M, Lukač Ž.. Camera and LiDAR sensor fusion for 3D object tracking in a collision
avoidance system. In: 2021 Zooming Innovation in Consumer Technologies Conference (ZINC); 2021; Novi Sad,
Serbia: IEEE. p. 198–202.
195. Truong LNH, Clay E, Mora OE, Cheng W, Singh M, Jia X. Rotated mask region-based convolutional neural
network detection for parking space management system. Transp Res Rec. 2023;2677(1):1564–81. doi:10.1177/
03611981221105066.
196. Zhang P, Xiong L, Yu Z, Fang P, Yan S, Yao J, et al. Reinforcement learning-based end-to-end parking for automatic
parking system. Sensors. 2019;19(18):3996. doi:10.3390/s19183996.
197. Wang L, Yu B, Chen C. Parking area data collection and scenario extraction for the purpose of automatic parking
ADAS function. In: IOP Conference Series: Materials Science and Engineering; 2020; Chengdu, China: IOP
Publishing. 32026 p.
198. Guo Y, Shi H. Automatic parking system based on improved neural network algorithm and intelligent image
analysis. Comput Intell Neurosci. 2021;2021(1):4391864. doi:10.1155/2021/4391864.
199. Rahim MA, Rahman MA, Rahman MM, Asyhari AT, Bhuiyan MZA, Ramasamy D. Evolution of IoT-enabled
connectivity and applications in automotive industry: a review. Veh Commun. 2021;27:100285. doi:10.1016/j.
vehcom.2020.100285.
200. Rahman MA, Rahim MA, Rahman MM, Moustafa N, Razzak I, Ahmad T, et al. A secure and intelligent framework
for vehicle health monitoring exploiting big-data analytics. IEEE Trans Intell Transp Syst. 2022;23(10):19727–42.
doi:10.1109/TITS.2021.3138255.
201. Rahim MA, Rahman MA, Rahman MM, Zaman N, Moustafa N, Razzak I. An intelligent risk management
framework for monitoring vehicular engine health. IEEE Trans Green Commun Netw. 2022;6(3):1298–306. doi:10.
1109/TGCN.2022.3179350.
202. Reiter G, Williams K. AI-based signal integrity monitoring for integrated vehicle health management. In: ELIV;
2021. Germany: VDI Wissensforum GmbH. p. 143–50.
203. Gomes IP, Wolf DF. Health monitoring system for autonomous vehicles using dynamic Bayesian networks for
diagnosis and prognosis. J Intell Robot Syst. 2021;101(1):19. doi:10.1007/s10846-020-01293-y.
204. Gupta N, Khosravy M, Patel N, Dey N, Crespo RG. Lightweight computational intelligence for IoT health
monitoring of off-road vehicles: enhanced selection log-scaled mutation GA structured ANN. IEEE Trans Ind
Inform. 2021;18(1):611–9. doi:10.1109/TII.2021.3072045.
3688 Comput Mater Contin. 2025;82(3)

205. Koricanac I. Impact of AI on the automobile industry in the U.S. SSRN Electron J. 2021. doi:10.2139/ssrn.3841426.
206. Tokody D, Ady L, Hudasi LF, Varga PJ, Hell P. Collaborative robotics research: subiko project. Procedia Manuf.
2020;46(2):467–74. doi:10.1016/j.promfg.2020.03.068.
207. Despot K, Srebrenkoska S, Sandeva V. The role of artificial intelligence in automotive design. KNOWLEDGE—Int
J. 2023;61(3):423–9.
208. Manimuthu A, Venkatesh VG, Shi Y, Sreedharan VR, Koh SCL. Design and development of automobile assembly
model using federated artificial intelligence with smart contract. Int J Prod Res. 2021;60(1):111–35. doi:10.1080/
00207543.2021.1988750.
209. Demlehner Q, Schoemer D, Laumer S. How can artificial intelligence enhance car manufacturing? A Delphi study-
based identification and assessment of general use cases. Int J Inf Manage. 2021;58:102317. doi:10.1016/j.ijinfomgt.
2021.102317.
210. Kinkel S, Baumgartner M, Cherubini E. Prerequisites for the adoption of AI technologies in manufacturing-
Evidence from a worldwide sample of manufacturing companies. Technovation. 2022;110(5):102375. doi:10.1016/j.
technovation.2021.102375.
211. Yue G, Tailai G, Dan W. Multi-layered coding-based study on optimization algorithms for automobile production
logistics scheduling. Technol Forecast Soc Change. 2021;170(6):120889. doi:10.1016/j.techfore.2021.120889.
212. Stauder M, Kühl N. AI for in-line vehicle sequence controlling: development and evaluation of an adaptive
machine learning artifact to predict sequence deviations in a mixed-model production line. Flexible Serv Manuf
J. 2022;34(3):709–47. doi:10.1007/s10696-021-09430-x.
213. Arunmozhi M, Venkatesh VG, Arisian S, Shi Y, Sreedharan VR. Application of blockchain and smart con-
tracts in autonomous vehicle supply chains: an experimental design. Transp Res Part E: Logist Transp Rev.
2022;165(3):102864. doi:10.1016/j.tre.2022.102864.
214. Soegoto ES, Utami RD, Hermawan YA. Influence of artificial intelligence in automotive industry. In: Journal of
Physics: Conference Series; 2019; Bristol, UK: IOP Publishing Ltd.
215. Mandala V. Revolutionizing automotive supply chain: enhancing inventory management with AI and machine
learning. Univ J Comput Sci Commun. 2024;3(1):10–22. doi:10.31586/ujcsc.2024.918.
216. Evangeliou N, Stamatis G, Bravos G, Plorin D, Stark D. AI for inbound logistics optimisation in automotive
industry. In: Artificial intelligence for digitising industry−applications. River Publishers; 2022. p. 11–9.
217. Borodavko B, Illés B, Bányai Á.. Role of artificial intelligence in supply chain. Acad J Manuf Eng. 2021;19(1):75–9.
218. Arena F, Collotta M, Luca L, Ruggieri M, Termine FG. Predictive maintenance in the automotive sector: a literature
review. Math Comput Appl. 2021;27(1):2. doi:10.3390/mca27010002.
219. Ersöz OÖ, İnal AF, Aktepe A, Türker AK, Ersöz S. A systematic literature review of the predictive maintenance
from transportation systems aspect. Sustainability. 2022;14(21):14536. doi:10.3390/su142114536.
220. Hossain MN, Rahman MM, Ramasamy D. Artificial intelligence-driven vehicle fault diagnosis to revolutionize
automotive maintenance: a review. Comput Model Eng Sci. 2024;141(2):951–96. doi:10.32604/cmes.2024.056022.
221. Li W, Li G, Kamarthi S. The study of trends in AI applications for vehicle maintenance through keyword co-
occurrence network analysis. Int J Progn Health Manag. 2023;14(2):1–17. doi:10.36001/ijphm.2023.v14i2.3583.
222. Theissler A, Pérez-Velázquez J, Kettelgerdes M, Elger G. Predictive maintenance enabled by machine learning: use
cases and challenges in the automotive industry. Reliab Eng Syst Saf. 2021;215(19):107864. doi:10.1016/j.ress.2021.
107864.
223. Aravind R, Shah CV. Physics model-based design for predictive maintenance in autonomous vehicles using AI.
Int J Sci Res Manag. 2023;11(9):932–46. doi:10.18535/ijsrm/v11i09.ec06.
224. Narayanan TSA, Padhy SC. Artificial intelligence for predictive maintenance of armoured fighting vehicles engine.
Indian J Artif Intell Neural Netw. 2023;3(5):1–12. doi:10.54105/ijainn.E1071.083523.
225. Wang Q, Wang H, Pan H. A constrained-time-based algorithm for vehicle maintain prediction. In: 5th Interna-
tional Conference on Information Science, Electrical and Automation Engineering (ISEAE 2023); 2023; Wuhan,
China: SPIE. p. 352–9.
Comput Mater Contin. 2025;82(3) 3689

226. Sharma N, Kalra M. Predictive maintenance for commercial vehicles tyres using machine learning. In: 2022
13th International Conference on Computing Communication and Networking Technologies (ICCCNT); 2022;
Kharagpur, India: IEEE. p. 1–6.
227. Chen C, Liu Y, Sun X, Di Cairano-Gilfedder C, Titmus S. An integrated deep learning-based approach for
automobile maintenance prediction with GIS data. Reliab Eng Syst Saf. 2021;216:107919. doi:10.1016/j.ress.2021.
107919.
228. Vasavi S, Aswarth K, Sai Durga Pavan T, Anu Gokhale A. Predictive analytics as a service for vehicle health
monitoring using edge computing and AK-NN algorithm. Mater Today: Proc. 2021;46(7):8645–54. doi:10.1016/j.
matpr.2021.03.658.
229. Biddle L, Fallah S. A novel fault detection, identification and prediction approach for autonomous vehicle
controllers using SVM. Autom Innov. 2021;4(3):301–14. doi:10.1007/s42154-021-00138-0.
230. Hu H, Luo H, Deng X. Health monitoring of automotive suspensions: a LSTM network approach. Shock Vib.
2021;2021(1):6626024. doi:10.1155/2021/6626024.
231. Mckinley T, Somwanshi M, Bhave D, Verma S. Identifying NOx sensor failure for predictive maintenance of diesel
engines using explainable AI. In: PHM Society European Conference; 2020. 11 p.
232. Krishnamurthy V, Nezafati K, Stayton E, Singh V. Explainable AI framework for imaging-based predictive
maintenance for automotive applications and beyond. Data-Enabled Discov Appl. 2020;4(1):7. doi:10.1007/s41688-
020-00042-2.
233. Voronov S, Krysander M, Frisk E. Predictive maintenance of lead-acid batteries with sparse vehicle operational
data. Int J Progn Health Manag. 2020;11(1):1–17. doi:10.36001/ijphm.2020.v11i1.2608.
234. Wang S, Liu Y, di Cairano-Gilfedder C, Titmus S, Naim MM, Syntetos AA. Reliability analysis for automobile
engines: conditional inference trees. Procedia CIRP. 2018;72(2):1392–7. doi:10.1016/j.procir.2018.03.065.
235. Shafi U, Safi A, Shahid AR, Ziauddin S, Saleem MQ. Vehicle remote health monitoring and prognostic maintenance
system. J Adv Transp. 2018;2018(1):8061514. doi:10.1155/2018/8061514.
236. Giobergia F, Baralis E, Camuglia M, Cerquitelli T, Mellia M, Neri A, et al. Mining sensor data for predictive
maintenance in the automotive industry. In: 2018 IEEE 5th International Conference on Data Science and
Advanced Analytics (DSAA); 2018; Turin, Italy: IEEE. p. 351–60.
237. Goikoetxea-Gonzalez J, Casado-Mansilla D, López-de-Ipiña D. The role of IoT devices in sustainable car
expenses in the context of the intelligent mobility: a comparative approach. Appl Sci. 2022;12(3):1080. doi:10.3390/
app12031080.
238. Jategaonkar A, Lenin NC. High-voltage electric water pump for the application of electric vehicle-modeling, design
and analysis. In: Kamaraj V, Ravishankar J, Jeevananthan S, editors. Emerging solutions for e-Mobility and smart
grids. Singapore: Springer; 2021. p. 201–12.
239. Morris J. Internal combustion cars will be an expensive luxury compared to electric by 2027 [Internet]. [cited
2022 Feb 26]. Available from: https://www.forbes.com/sites/jamesmorris/2021/05/15/internal-combustion-cars-
will-be-an-expensive-luxury-compared-to-electric-by-2027/?sh=2782225e5189.
240. EVs to be cheaper to produce than fossil fuel vehicles by 2027 ∣ Daily Sabah [Internet]. [cited 2022
Feb 26]. Available from: https://www.dailysabah.com/business/automotive/evs-to-be-cheaper-to-produce-than-
fossil-fuel-vehicles-by-2027.
241. Mustapa SI, Ayodele BV, Ishak WWM, Ayodele FO. Evaluation of cost competitiveness of electric vehicles in
Malaysia using life cycle cost analysis approach. Sustainability. 2020;12(13):5303. doi:10.3390/su12135303.
242. Fechtner H, Schmuelling B. Improved range prediction for electric vehicles by a smart tire pressure monitoring
system. In: 2018 IEEE Conference on Technologies for Sustainability (SusTech); 2018; Long Beach, CA, USA. p.
1–8.
243. Selvabharathi D, Muruganantham N. Experimental analysis on battery based health monitoring system for electric
vehicle. Mater Today: Proc. 2021;45(3):1552–8. doi:10.1016/j.matpr.2020.08.303.
244. Zhang Y, Li T, Yan X, Wang L, Zhang J, Diao X, et al. Electric vehicle charging fault monitoring and warning method
based on battery model. World Electr Veh J. 2021;12(1):14. doi:10.3390/wevj12010014.
3690 Comput Mater Contin. 2025;82(3)

245. Wang C, Ji T, Mao F, Wang Z, Li Z. Prognostics and health management system for electric vehicles with a
hierarchy fusion framework: concepts, architectures, and methods. Adv Civil Eng. 2021;2021(1):1–11. doi:10.1155/
2021/6685900.
246. Blazevic R, Veledar O, Stolz M, Macher G. Toward safety-critical artificial intelligence (AI)-based embedded
automotive systems. SAE Int J Connected Automated Veh. 2024;8(12–08-01–0007). doi:10.4271/12-08-01-0007.
247. Puertas CAP, Galhardi AC. Integrating ergonomic and artificial intelligence in the automotive. In: SAE technical
paper; 2024.
248. Castaño F, Strzelczak S, Villalonga A, Haber RE, Kossakowska J. Sensor reliability in cyber-physical systems using
internet-of-things data: a review and case study. Remote Sens. 2019;11(19):2252. doi:10.3390/rs11192252.
249. Yeong DJ, Velasco-Hernandez G, Barry J, Walsh J. Sensor and sensor fusion technology in autonomous vehicles: a
review. Sensors. 2021;21(6):2140. doi:10.3390/s21062140.
250. Velusamy SP, Ghannam MY, Kadry HM. Automotive sensor infrastructure-challenges and opportunities. In: 2022
IEEE International Symposium on Circuits and Systems (ISCAS); 2022; Austin, TX, USA: IEEE. p. 1018–22.
251. Tubaro P, Casilli AA. Micro-work, artificial intelligence and the automotive industry. J Ind Bus Econ.
2019;46(3):333–45. doi:10.1007/s40812-019-00121-1.
252. Wasner C, Traxler J. AI and automotive safety. ATZelectron Worldwide. 2019;14(11):50–3. doi:10.1007/s38314-019-
0128-z.
253. Jain M, Kulkarni P. Application of AI, IOT and ML for business transformation of the automotive sector. In: 2022
International Conference on Decision Aid Sciences and Applications (DASA); 2022; Chiangrai, Thailand: IEEE. p.
1270–5.
254. Zhang R, Zhao W, Wang C. Research on coordinated control of braking mode switching for enhancing
switching quality considering communication delay and time-varying model parameters. IEEE Trans Intell Veh.
2022;8(3):2494–513. doi:10.1109/TIV.2022.3225403.
255. Grigorescu S, Cocias T, Trasnea B, Margheri A, Lombardi F, Aniello L. Cloud2edge elastic AI framework for
prototyping and deployment of AI inference engines in autonomous vehicles. Sensors. 2020;20(19):5450. doi:10.
3390/s20195450.
256. Thamizhazhagan P, Sujatha M, Umadevi S, Priyadarshini K, Parvathy VS, Pustokhina IV, et al. AI based traffic flow
prediction model for connected and autonomous electric vehicles. Comput Mater Contin. 2022;70(2):3333–47.
doi:10.32604/cmc.2022.020197.
257. Siddiqui F, Khan R, Sezer S. Bird’s-eye view on the automotive cybersecurity landscape & challenges in adopting
AI/ML. In: 2021 Sixth International Conference on Fog and Mobile Edge Computing (FMEC); 2021; Gandia, Spain:
IEEE. p. 1–6.
258. Aldhyani THH, Alkahtani H. Attacks to automatous vehicles: a deep learning algorithm for cybersecurity. Sensors.
2022;22(1):360. doi:10.3390/s22010360.
259. Török Á., Szalay Z, Sághi B. New aspects of integrity levels in automotive industry—cybersecurity of automated
vehicles. IEEE Trans Intell Transp Syst. 2020;23(1):383–91. doi:10.1109/TITS.2020.3011523.
260. Morris D, Madzudzo G, Garcia-Perez A. Cybersecurity threats in the auto industry: tensions in the knowledge
environment. Technol Forecast Soc Change. 2020;157(1):120102. doi:10.1016/j.techfore.2020.120102.
261. Eze CP, Emmanuel J, Onietan CIO, Isewon I, Oyelade J. Systematic review on the recent trends of cybersecurity
in automobile industry. In: 2023 International Conference on Science, Engineering and Business for Sustainable
Development Goals (SEB-SDG); 2023; Omu-Aran, Nigeria: IEEE. p. 1–7.
262. Jadoon AK, Wang L, Li T, Zia MA. Lightweight cryptographic techniques for automotive cybersecurity. Wirel
Commun Mob Comput. 2018;2018(1):1640167. doi:10.1155/2018/1640167.
263. Tiberti W, Civino R, Gavioli N, Pugliese M, Santucci F. A hybrid-cryptography engine for securing intra-vehicle
communications. Appl Sci. 2023;13(24):13024. doi:10.3390/app132413024.
264. Kumar A, Gholve A, Kotalwar K. Automotive security solution using hardware security module (HSM). In: SAE
technical paper; 2024.
265. Baldini G, Giuliani R, Gemo M, Dimc F. On the application of sensor authentication with intrinsic physical features
to vehicle security. Comput Electr Eng. 2021;91(2):107053. doi:10.1016/j.compeleceng.2021.107053.
Comput Mater Contin. 2025;82(3) 3691

266. Luo F, Zhang X, Yang Z, Jiang Y, Wang J, Wu M, et al. Cybersecurity testing for automotive domain: a survey.
Sensors. 2022;22(23):9211. doi:10.3390/s22239211.
267. Lütge C, Poszler F, Acosta AJ, Danks D, Gottehrer G, Mihet-Popa L, et al. AI4People: ethical guidelines for the auto-
motive sector-fundamental requirements and practical recommendations. Int J Technoethics. 2021;12(1):101–25.
doi:10.4018/IJT.
268. Wang H, Khajepour A, Cao D, Liu T. Ethical decision making in autonomous vehicles: challenges and research
progress. IEEE Intell Transp Syst Mag. 2020;14(1):6–17. doi:10.1109/MITS.2019.2953556.
269. Sharma A, Gupta A, Bhargava A, Rawat A, Yadav P, Gupta D. From Sci-Fi to reality: the evolution of human—
computer interaction with artificial intelligence. In: 2023 2nd International Conference on Applied Artificial
Intelligence and Computing (ICAAIC); 2023; Salem, India: IEEE. p. 127–34.
270. Hussain R, Zeadally S. Autonomous cars: research results, issues, and future challenges. IEEE Commun Surv
Tutorials. 2018;21(2):1275–313. doi:10.1109/COMST.2018.2869360.
271. Muhammad K, Ullah A, Lloret J, Del Ser J, de Albuquerque VHC. Deep learning for safe autonomous driving:
current challenges and future directions. IEEE Trans Intell Transp Syst. 2020;22(7):4316–36. doi:10.1109/TITS.2020.
3032227.
272. Grigorescu S, Trasnea B, Cocias T, Macesanu G. A survey of deep learning techniques for autonomous driving. J
Field Robot. 2020;37(3):362–86. doi:10.1002/rob.21918.
273. Ma Y, Wang Z, Yang H, Yang L. Artificial intelligence applications in the development of autonomous vehicles: a
survey. IEEE/CAA J Automatica Sinica. 2020;7(2):315–29. doi:10.1109/JAS.2020.1003021.
274. Bathla G, Bhadane K, Singh RK, Kumar R, Aluvalu R, Krishnamurthi R, et al. Autonomous vehicles and intelligent
automation: applications, challenges, and opportunities. Mob Inf Syst. 2022;2022(1):7632892. doi:10.1155/2022/
7632892.
275. Gao X, Bian X. Autonomous driving of vehicles based on artificial intelligence. J Intell Fuzzy Syst.
2021;41(4):4955–64. doi:10.3233/JIFS-189982.
276. Sharma M, Sharma M, Sharma M, Yadav KS, Shukla S. Industry 4.0 technologies for smart manufacturing: a
systematic review of machine learning methods for predictive maintenance. In: 2023 International Conference on
Self Sustainable Artificial Intelligence Systems (ICSSAS); 2023; Erode, India. p. 397–403.
277. Jerry A, Madrid. The role of artificial intelligence in automotive manufacturing and design. Int J Adv Resn Sci,
Commun Technol. 2023;3(1):798–802.
278. Lee J, Chua PC, Chen L, Ng PHN, Kim Y, Wu Q, et al. Key enabling technologies for smart factory in automotive
industry: status and applications. Int J Precision Eng Manuf—Smart Technol. 2023;1(1):93–105. doi:10.57062/ijpem-
st.2022.0017.
279. Ajoudani A, Zanchettin AM, Ivaldi S, Albu-Schäffer A, Kosuge K, Khatib O. Progress and prospects of the human—
robot collaboration. Auton Robots. 2018;42(5):957–75. doi:10.1007/s10514-017-9677-2.
280. Othman U, Yang E. Human-robot collaborations in smart manufacturing environments: review and outlook.
Sensors. 2023;23(12):5663. doi:10.3390/s23125663.
281. Bogue R. The changing face of the automotive robotics industry. In Robot: Int J Robot Res Appl. 2022;49(3):386–90.
doi:10.1108/IR-01-2022-0022.
282. Michalos G, Kousi N, Karagiannis P, Gkournelos C, Dimoulas K, Koukas S, et al. Seamless human robot
collaborative assembly—an automotive case study. Mechatronics. 2018;55:194–211. doi:10.1016/j.mechatronics.2018.
08.006.
283. de Costa GM, Petry MR, Moreira AP. Augmented reality for human-robot collaboration and cooperation in
industrial applications: a systematic literature review. Sensors. 2022;22(7):2725. doi:10.3390/s22072725.
284. Razdan R, Akbaş Mİ., Sell R, Bellone M, Menase M, Malayjerdi M. Polyverif: an open-source environment for
autonomous vehicle validation and verification research acceleration. IEEE Access. 2023;11(3):28343–54. doi:10.
1109/ACCESS.2023.3258681.
285. Ma Y, Sun C, Chen J, Cao D, Xiong L. Verification and validation methods for decision-making and planning of
automated vehicles: a review. IEEE Trans Intell Veh. 2022;7(3):480–98. doi:10.1109/TIV.2022.3196396.
3692 Comput Mater Contin. 2025;82(3)

286. Ebadinezhad S, Stanley KA, Engo GM, Osemeha NK, Mayar AS. The impact of AI on traffic management and
safety in the internet of vehicles through mobile edge computing. In: 2024 International Conference on Inventive
Computation Technologies (ICICT); 2024; Lalitpur, Nepal: IEEE. p. 319–26.
287. Saleh SN, Fathy C. A novel deep-learning model for remote driver monitoring in SDN-based internet of
autonomous vehicles using 5G technologies. Appl Sci. 2023;13(2):875. doi:10.3390/app13020875.
288. You X, Zhang C, Tan X, Jin S, Wu H. AI for 5G: research directions and paradigms. Sci China Inf Sci.
2019;62(2):1–13. doi:10.1007/s11432-018-9596-5.
289. Kim H, Ben-Othman J, Mokdad L, Son J, Li C. Research challenges and security threats to AI-driven 5G virtual
emotion applications using autonomous vehicles, drones, and smart devices. IEEE Netw. 2020;34(6):288–94.
doi:10.1109/MNET.011.2000245.
290. Guo H, Zhou X, Liu J, Zhang Y. Vehicular intelligence in 6G: networking, communications, and computing. Veh
Commun. 2022;33:100399. doi:10.1016/j.vehcom.2021.100399.
291. Rudin C. Stop explaining black box machine learning models for high stakes decisions and use interpretable models
instead. Nat Mach Intell. 2019;1(5):206–15. doi:10.1038/s42256-019-0048-x.
292. Moreira C, Chou YL, Velmurugan M, Ouyang C, Sindhgatta R, Bruza P. LINDA-BN: an interpretable probabilistic
approach for demystifying black-box predictive models. Decis Support Syst. 2021;150:113561. doi:10.1016/j.dss.2021.
113561.
293. Lakkaraju H, Arsov N, Bastani O. Robust and stable black box explanations. In: 37th International Conference on
Machine Learning; 2020; PMLR. p. 5628–38.