International Journal of Heavy Vehicle Systems (31 papers in press)

Regular Issues

• Tyre modelling influence on Performance Base Standard (PBS) performance  
  by Tokologo Komana, Schalk Els, Carl Martin Becker 
  Abstract: Initially Performance Base Standard (PBS) required that assessments are conducted with models of the actual tyres the vehicle will operate with. This led to practical difficulties for both assessors and operators. Assessors found it difficult to source accurate tyre experimental data and operators found restricting the vehicle to a set of tyres negatively affects their business case. The PBS tyre review resolved that all PBS assessment must be conducted with this generic tyre, thus tyre is treated as a test condition rather than a model parameter. While this resolution is pragmatic for both operators and assessors, cornering stiffness is known to significantly influence vehicle handling. So, this study investigates the influence of corning stiffness on PBS performance. Results show that tyres with low cornering stiffnesses significantly influence the PBS performance ranging between level 1-4. Tyres with a high cornering stiffness do not significantly influence the PBS performance of heavy vehicles.
  Keywords: Performance Base Standard (PBS); cornering stiffness; Pacejka tyre model; Rearward Amplification (RA); High Speed Transient Offtracking (HSTO); Yaw Damping Coefficient (YDC).
  
  
• The effect of piston pin offset on piston friction in heavy vehicle compressors: a comparative FEA study of conventional and new crank-connecting rod mechanisms  
  by Ozgur Cetin, Melih Okur 
  Abstract: In crank-connecting rod mechanisms, a major portion of mechanical losses results from piston-cylinder friction, primarily caused by the piston’s secondary motion. One of the most influential parameters affecting this motion is the piston pin offset. This study introduces a novel crank-connecting rod mechanism with reduced lateral thrust as an alternative to the conventional design. The effects of various pin offsets on lateral piston forces were analyzed using 3D finite element analysis. A two-stage air brake compressor model, commonly used in heavy-duty vehicles, was employed along with thermodynamic assumptions. Results showed that offsets applied toward the anti-thrust side reduced lateral forces in both systems. At -3 mm offset, the novel mechanism produced 81% less friction compared to the conventional one. It was also concluded that to avoid eccentric friction caused by piston moments, the pin offset should be kept within a limited range.
  Keywords: heavy vehicle compressor; secondary movement of piston; piston pin offset; piston friction; FEA; finite element analysis.
  DOI: 10.1504/IJHVS.2025.10074970
   
  
• Research on following braking control of the aircraft engine-off taxi towing system  
  by Kai Qi, Juanjuan Wei 
  Abstract: In traditional aircraft towing, only the towbarless towing vehicle (TLTV) provides braking, which poses risks at higher speeds due to the aircraft's large inertia. The new-generation aircraft engine-off taxi towing system (AEOTTS) eliminates the need for a TLTV driver, allowing the pilot alone to control braking. In this mode, the TLTV follows the aircraft based on force and motion signals from the nose wheel. During braking, the combined effort from both the aircraft and the TLTV shortens the braking distance and keeps the nose landing gear (NLG) load within a safe limit. A coupled vertical-longitudinal model was developed, along with a validated tire dynamics model. A fuzzy PID slip rate controller was designed for following braking. Co-simulations in Adams/View and Matlab/Simulink demonstrated that this strategy significantly reduces both braking distance and peak NLG load, enhancing towing safety.
  Keywords: AEOTTS; aircraft engine-off taxi towing system; TLTV; towbarless towing vehicle; braking dynamic; slip rate control.
  DOI: 10.1504/IJHVS.2025.10075065
   
  
• Reconfigurable multi-unit, multi-axle, multi-articulation heavy vehicle (RUAAHV) driver modelling for PBS assessments  
  by Tokologo Komana, Schalk Els, Herman A. Hamersma 
  Abstract: Performance Based Standards (PBS) is a new paradigm in heavy vehicle regulation based on the vehicle’s safety performance, rather than on the vehicle’s physical attributes such as dimensions and mass. Commonly, PBS assessments through simulations are preferred, because field testing is more costly and poses a safety risk to the test engineers and equipment. Assessing PBS performance through simulation requires a robust driver model to steer the simulation model to perform the different PBS manoeuvres. This study develops a reconfigurable driver model in Simulink to steer an Adams vehicle model through co-simulation. Results show that the driver modelling approach achieves the PBS path following error requirements of 0.05 m and 0.03 m for low speed and high speed manoeuvres, respectively. The reconfigurable approach makes the driver model adaptable to various vehicle configurations (five different vehicles are simulated in this study) and various speeds (5-90 km/h).
  Keywords: heavy vehicle; PBS; performance based standards; LQR; linear quadratic regulator; Kalman filter (KF); driver model; B-double; truck/trailer; Adams view.
  DOI: 10.1504/IJHVS.2025.10075112
   
  
• Electric powertrain sizing for light cargo vehicles in developing markets  
  by Sriniket Chavan, Satyajit Patil 
  Abstract: Tailpipe emissions from automobiles are one of the significant sources of air pollution, harming the environment. While electric vehicles are known for environmental friendliness, their penetration in cargo transportation is still questionable; thus, it is proposed to electrify the powertrain of a typical cargo vehicle. The battery capacity and electric motor rating are the critical decisions for an electric powertrain. These influence vehicle performance regarding range, top speed, and acceleration. This work presents a methodology to develop an electric powertrain using a modelling and simulation approach. The analytical approach estimated the battery sizing and motor rating, while the simulation studies validated the estimation for the FTP 75 drive cycle. The simulation results indicate a battery capacity of 96 kWh with a motor rating of 30 kW to meet the performance demands of the cargo vehicle. The results could serve as benchmark sizing for further electrification development projects.
  Keywords: electric vehicles; emission; modelling; simulation; battery sizing; electric powertrain.
  DOI: 10.1504/IJHVS.2025.10075195
   
  
• Research on vehicle dynamic weighing method based on narrow strip strain sensor  
  by Zicheng Qi, Jianyun Shen, Xuyang Xu, Ruolan Wang 
  Abstract: To address dynamic weighing errors in vehicles caused by speed variations and uneven road surfaces, a novel approach combines trapezoidal-normal distribution convolution modeling with dual vertical narrow-strip sensor arrays. This layout minimizes road-induced fluctuations, while a ladder function simulates driving state changes (acceleration/deceleration, grounding length). A normal distribution-based transfer function characterizes sensor dynamics. Comparative tests demonstrate superior accuracy (0.78% average error) and stability versus traditional averaging methods. The system maintains precision across speeds and axle counts, enabling reliable freight vehicle weighing with enhanced robustness to operational variables. This integration of mathematical modelling and optimised sensor placement provides an effective solution to resolve uncertainties in dynamic weighing under real-world scenarios.
  Keywords: normally distributed response function; narrow strip sensor; dynamic weighing; tire pressure model.
  DOI: 10.1504/IJHVS.2025.10075460
   
  
• Modelling and validation of the lateral dynamics of multi-trailer articulated heavy vehicles for active safety systems design and optimisation  
  by Shenjin Zhu, Yuping He 
  Abstract: Active safety systems (ASSs) are promising in enhancing the lateral dynamics of multi-trailer articulated heavy vehicles (MTAHVs). One essential technique for ASS development is model-based controller design, which can predict and control the lateral dynamics of MTAHVs under various operating conditions. To discuss the modelling of MTAHVs, an overview of state-of-theart techniques for tyre and vehicle modelling is presented. An MTAHV with the configuration of B-train double is selected for this study. To design, coordinate and optimise conventional ASSs for MTAHVs, four typical models, i.e., linear yaw-plane, linear yaw-roll, 2-Dimensional (2D) nonlinear double-track, and 3-dimensional (3D) nonlinear EoM, are derived and validated using a 3D nonlinear TruckSim model. According to their complexities and predicting capabilities evaluated in simulations under low and high lateral acceleration operations, these models are recommended for respective applications to ASS designs.
  Keywords: multi-trailer articulated heavy vehicles; active safety systems; linear yaw-plane model; linear yaw-roll model; nonlinear yaw-plane model; nonlinear EoM yaw-roll model; TruckSim model; model allocatio.
  DOI: 10.1504/IJHVS.2025.10076144
   
  
• Modelling of rolling resistance of agricultural tyre based on machine learning algorithms  
  by Ergün Çıtıl, Kazım Çarman, Alper Taner 
  Abstract: Axle loads on tractor tyres increase proportionally with tractor power, making rolling resistance an performance parameter. This study investigates the rolling resistance of tractor drive tyres under controlled conditions using a single-wheel test rig in a soil chamber and evaluates five machine learning models. Experiments were conducted at vertical loads of 3.56.5 kN, tyre inflation pressures of 150240 kPa, and a constant speed of 0.45 m/s. Results showed that increasing inflation pressure reduced the tyre soil contact area, whereas higher vertical loads enlarged it. Rolling resistance ranged from 240 N to 3170 N and showed strong correlations with both vertical load and tyre inflation pressure. Support vector regression, multi-layer perceptron regressor, extreme gradient boosting regressor, k-nearest neighbours regressor, and random forest regressor were used for prediction. Among these models, the multi-layer perceptron regressor achieved the best performance, with a mean absolute percentage error of 7.66% and a correlation coefficient of 0.98.
  Keywords: agricultural tyre; machine learning algorithms; rolling resistance; tyre inflation pressure; tyre contact surface area.
  DOI: 10.1504/IJHVS.2025.10076145
   
  
• Study on lane changing trajectory planning and tracking control of semitrailers based on multi-objective optimisation  
  by Xingkun Li, Aihong Meng, Yawei Li, Yushuai Zhao, Guangyu Tian, Charles A. Garris 
  Abstract: With the rapid advancement of autonomous driving, intelligent semitrailers present significant development opportunities but also unique challenges, including large frontrear wheel track differences during turning, significant load-dependent mass variation, and high sensitivity to fuel economy. In this study, a six-axle, three-degree-of-freedom dynamic model of a semitrailer is established, and an engine fuel consumption model is developed through hardware-in-the-loop testing. A multi-objective lane-changing trajectory optimisation method based on a fifth-order polynomial and genetic algorithm is proposed, considering lane-changing efficiency, comfort, and fuel economy. An integrated lateral controller combining LQR, single-point preview, and feedforward control is designed to track the optimised trajectory and speed. Co-simulation results using TruckSim and Simulink demonstrate that incorporating fuel economy into trajectory planning improves safety, stability, and fuel efficiency, highlighting its significance for commercial vehicles.
  Keywords: semitrailer; trajectory planning; tracking control; genetic algorithm; fuel economy.
  DOI: 10.1504/IJHVS.2025.10076146
   
  
• Design of dynamic zero position for single-pedal motor torque considering driver braking habits  
  by Xuhao Zhang, Yufang Li, Yuhang Wang, Jihang Li, Siyu Xu, Dexin Gao, Tianci Zhang 
  Abstract: Single-pedal control relies on defining the mapping between pedal position and motor zero torque output. The zero torque point determines the pedals drive/brake threshold, and its speed-dependent curve, known as the zero torque line, critically affects vehicle performance. Existing systems use fixed zero torque lines to separate drive and brake torque for energy recovery, yet they cannot simultaneously optimise energy recovery and acceleration response. To resolve this trade-off, this paper introduces a dynamic zero torque position strategy. The proposed method adjusts the zero torque point in real time, maintaining high energy recovery efficiency while preserving driving comfort. Simulations under urban conditions demonstrate that the dynamic scheme outperforms conventional fixed-line approaches in both adaptability and total energy recovered.
  Keywords: single-pedal control technology; pedal position; motor zero-torque; energy recovery; urban driving conditions.
  DOI: 10.1504/IJHVS.2025.10076253
   
  
• Development and verification of an adaptive ECMS incorporating efficiency factors for autonomous driving range-extended hybrid mining trucks  
  by Tao Li, Zhao Zhiguo, Huiyong Chen, Jianyu Yang, Peihong Shen 
  Abstract: Considering that range-extended hybrid mining trucks cannot be externally charged during operation, this paper proposes an Adaptive Equivalent Consumption Minimization Strategy (A-ECMS) incorporating efficiency factors. Firstly, a reference SOC trajectory planning method based on Dynamic Programming (DP) was designed. The operation was divided into four modes based on load status and SOC.Secondly, an ECMS control strategy incorporating efficiency factors is proposed to quantify the efficiency loss of battery charging and discharging. Genetic Algorithm (GA) is employed for optimization to obtain the optimal equivalent factor and efficiency factors under each SOC mode. Finally, real-world testing validated that the proposed ECMS operates in real time with strong adaptability. Compared to the rule-based strategy and conventional ECMS, the proposed strategy achieved 3.24% and 2.41% reductions in equivalent fuel consumption, The engine operating point loss ratio and battery power loss ratio decreased to 2.43% and 1.51%, respectively.
  Keywords: range-extended hybrid mining trucks; efficiency factor; fuel consumption equivalent factor; A-ECMS; adaptive equivalent consumption minimisation strategy; energy management strategy.
  DOI: 10.1504/IJHVS.2025.10076383
   
  
• Military unmanned ground vehicles: technologies, capabilities, and future trends  
  by Hossam Ragheb 
  Abstract: Military unmanned ground vehicles have evolved from experimental platforms to critical assets in reconnaissance, mine clearance, logistics, and combat support missions. This review summarizes the historical development, key enabling technologies, operational applications, and current challenges. We examine the origins of AI-driven autonomy from the 1990s to the present day, analyzing sensor fusion architectures, navigation algorithms, mobility solutions, and human-robot interaction paradigms. Technical challenges related to terrain sensing, autonomy integration, reliability, and multiple robot coordination are discussed in detail. We identify critical research directions, such as integrated decision-making structures, collaborative mobility sensing, swarm coordination, and ethical frameworks for killer autonomy. The review presents the cutting edge in military UGV technology and shows the way forward for the next generation of researchers, military planners, and policy makers.
  Keywords: UGVs; unmanned ground vehicles; military robotics; autonomous navigation; combat robotics; human-robot interaction; artificial intelligence; multi-robot systems.
  DOI: 10.1504/IJHVS.2026.10076914
   
  
• Intelligent suspension control system for autonomous vehicles based on multi-sensor information fusion  
  by Peng Ding, Wang Zhong, Gu Xiaoyong, Zhang Meijuan 
  Abstract: An intelligent suspension control method based on multi-sensor information fusion is proposed to enhance the safety and comfort of autonomous vehicles on damaged roads. A quarter-suspension vibration model integrating multi-sensor data is established to reveal the relationship between road roughness and vehicle vibration. A camera and radar are employed to scan uneven road conditions, developing a mathematical road roughness model. Information fusion and matching are achieved through edge intersection ratio detection and the global nearest neighbours (GNN) algorithm, ensuring high model reliability in complex environments. The optimal damping ratio is calculated using vehicle speed and road roughness data, enabling real-time suspension adaptation to road variations. Test results demonstrate that the maximum vibrational acceleration of the proposed system is reduced by over 43% compared to passive suspension, confirming the effectiveness of this intelligent control approach.
  Keywords: driverless vehicle; intelligent suspension control; multi-sensor information fusion; road roughness estimation; GNN algorithm; semi-active suspension; dynamical model; vibration control; road profile recognition.
  DOI: 10.1504/IJHVS.2025.10077061
   
  
• Enhancing heavy-duty natural gas engine performance and emissions prediction with diffractive deep neural networks  
  by Jayasheel Kumar Kalagatoor. Archakam, Rakesh Chandrashekar, Santosh Kumar Balraj, Rama Devi Chellappan 
  Abstract: This manuscript presents a comprehensive analysis of a heavy-duty natural gas spark ignition (SI) engine. The diffractive deep neural network (DDNN) is applied to predict key engine parameters such as peak pressure location, peak cylinder pressure, ignition lag, engine thermal efficiency, nitrogen oxides concentration, indicated mean effective pressure, critical angle, exhaust gas temperature, and combustion duration. The study includes a detailed configuration of the dynamometer test system used to measure torque, speed, fuel consumption, and emissions. It also outlines the conversion of a diesel engine to a natural gas SI engine and the associated modifications and benefits, including lower greenhouse gas emissions and reduced fuel costs. The SI engine thermodynamic cycle processes are modeled using differential equations to simulate engine behaviour. The DDNN model, implemented in Python, is evaluated using RMSE, MRE, and R
  Keywords: diesel engine; DDNN; diffractive deep neural network; compression; natural gas; temperature indicator; low emission; thermal efficiency.
  DOI: 10.1504/IJHVS.2025.10077667
   
  
• Development of a new prototype of dual drivetrain switching mechanism for tracked vehicles: modelling approaches  
  by Muhammad Luqman Abd. Rahman, Noor Hafizah Amer, Khisbullah Hudha, Zulkiffli Abd Kadir, Mohamed Ishak Saiddi Ali Firdaus, Syed Mohd Fairuz Syed Mohd Dardin 
  Abstract: Tracked vehicles provide higher capabilities in performance and mobility due to the effectiveness of track system. However, due to the construction of track system, it damages the surface of asphalt and concrete roads. To overcome this problem, a new dual drivetrain switching mechanism is developed for providing both wheels and tracks on various terrains. The new dual drivetrain switching mechanism consists of four movable arms with wheels where those arms are actuated via power screw mechanism. The new dual drivetrain switching mechanism developed using CAD model have been verified using non-parametric modeling where a 4th order polynomial model is used to follow the behavior of CAD model. Results show that the polynomial model can follow its behaviour with average percentage of deviation around 0.0152% to 4.1253%.
  Keywords: tracked vehicle; dual mode vehicle; wheel-cum-tracked vehicle; dual drivetrain system; hybrid wheel-track vehicle; switching mechanism; power screw actuation; terrain adaptability; mathematical modeling; CAD verification; movable arm kinematics; road surface preservation; polynomial regression.
  DOI: 10.1504/IJHVS.2025.10077786
   
  
• Severity Analysis of Railway Track Geometry Defects  
  by Sharad Nigam, Divya Kumar, Suryakant Shastri 
  Abstract: The railway train is a chariot that gives passengers a feeling of rapidity and safety Rapidity in transportation is much needed for the fast development of a country and the overall economy But not at the cost of the lives of passengers/public The safety of the passengers is a prominent goal of the railway authorities, who provide safety facilities and regular maintenance of the concerned components Railway tracks are one of the major components responsible for railway accidents mostly derailments A minor or major defect in a rail may cause accidents Various types of defects are detected on the rail track This paper is only concerned with geometry-type defects in the track This paper presents an analytical study of three types of geometry defects i e SURFACE, DIP, and XLEVEL to analyze defect severity and maintenance requirements A dataset is taken for the defect analysis from RAS Track Geometry Analytics.
  Keywords: geometry defect; SURFACE; XLEVEL; DIP; time gap.
  DOI: 10.1504/IJHVS.2026.10077919
   
  
• DDQN-based trajectory tracking control for unmanned ground vehicles: design and experimental validation  
  by Islam Hassan, Tamer Attia, A.M. Sharaf, Hossam Ragheb 
  Abstract: In this paper, geometric controllers are used to train Stanley techniques. To improve the tracking accuracy, the basic principle of the Stanley method, which works as a guide for the unmanned ground vehicle (UGV) to follow a predefined trajectory, is used. Then, the outputs of the controller are compensated using deep reinforcement learning (DRL) to correct the control inputs to follow the trajectory with high accuracy. The double deep Q-network (DDQN) algorithm is used for the training process to obtain the optimal policy of the proposed control framework. Comparative MATLAB and Simulink simulations and experimental tests were performed to confirm the effectiveness of the proposed control framework for accurate monitoring. Simulation results show that using the Stanley-based DDQN algorithm has higher tracking accuracy and smoothness than a conventional Stanley. The real UGV results are given to compare the Stanley and the proposed approach and show the higher trajectory tracking accuracy.
  Keywords: Stanley-RL; UGV; unmanned ground vehicle; DNN; deep neural network; DRL; deep reinforcement learning; trajectory tracking.
  DOI: 10.1504/IJHVS.2026.10078074
   
  
• Investigation of braking force distribution system for the pitch risk of heavy-duty vehicles  
  by Xiaopeng Yang, Zengbin Wu, Yong Zhang, He Gou, Liyu Yang, Shikai Wang, Wupeng Liu, Wei Zhou, Tongyue Zhang, Sen Xiao 
  Abstract: As the complexity of vehicle loads and driving environments increases, the performance requirements for braking systems become increasingly stringent. This study aims to achieve optimal braking performance for heavy-duty vehicles and reduce the risk of longitudinal tilt during braking by designing a brake force distribution control system. An in-depth analysis of the dynamics principles of wheel braking is performed. The designed system employs a closed-loop strategy on the basis of whole-vehicle deceleration. Simulation results show that the control system can dynamically adjust the brake force distribution in accordance with real-time conditions and fully utilize the maximum adhesion for braking. The goal of reducing the risk of longitudinal tilt and enhancing the vehicle’s operational stability during the braking process is thus achieved.
  Keywords: HDVs; heavy-duty vehicles; longitudinal tilt; hydraulic control valves; brake force distribution.
  DOI: 10.1504/IJHVS.2025.10078249
   
  
• Multi-mode switching control and experimental validation of an active hydro-pneumatic suspension for mining dump trucks  
  by Ziyan Zhao, Huiqing Yan, Yumeng Li, Anxin Sun, Fangwei Xie 
  Abstract: The conventional passive hydro-pneumatic suspensions on mining dump trucks exhibit limited adaptability, leading to compromised ride comfort under complex and impact-prone mining terrains. To address this, an active hydro-pneumatic suspension configuration capable of switching between four distinct stiffness-damping modes is proposed. The core of this study lies in a multi-mode switching control strategy predicated on known road profiles. This strategy incorporates a pre-processing stage for the road input, effectively implementing a delay compensation mechanism to counteract the inherent response lag of electro-hydraulic directional valves, thereby ensuring timely actuation. Moreover, the critical control step size is systematically optimised using a Genetic Algorithm, automating the parameter tuning process to enhance performance. The fidelity of the established nonlinear system model is rigorously validated through experimental characterisation of a physical prototype. Simulation results, grounded in this validated model, demonstrate that the proposed active system achieves a superior trade-off, significantly improving ride comfort metrics compared to its passive counterpart. This work provides a credible and model-verified framework for the development of highperformance active suspensions in heavy-duty vehicles.
  Keywords: active hydro-pneumatic suspension; multi-mode control; genetic algorithm; mining dump truck; ride comfort; experimental validation.
  DOI: 10.1504/IJHVS.2026.10078370
   
  
• Malware variant traffic identification using Jensen-Shannon quantum dilated convolutional neural network  
  by Arumugam Gonda, Srinath Doss, Kamal Kant Hiran 
  Abstract: The widespread adoption of the Internet of Things across diverse domains has led to a surge in malware variants, resulting in substantial damage. Building strong models and intrusion detection tools is essential to recognize and investigate cyberattacks. To tackle malware traffic detection in IoT-edge gateways, this paper introduces the Jensen-Shannon Quantum Dilated Convolutional Neural Network (JS-QDNN). Initially, an IoT edge gateway is simulated, followed by passing the log file to the feature extraction stage, where both network-level and traffic-based features are derived. Afterward, oversampling is employed for data augmentation, and malware variant traffic is subsequently recognized using JS-QDCNN, a hybrid model combining Quantum Dilated Convolutional Neural Network and Jensen-Shannon similarity and the detected malware is classified as Malware A and Malware B. Moreover, the JS-QDCNN computed 91.865% accuracy, 91.744% precision and 93.090% recall.
  Keywords: IoT; Internet of Things; deep learning; cyber threat; malicious traffic; malware variant traffic identification.
  DOI: 10.1504/IJHVS.2025.10078372
   
  
• Optimization of hydro-pneumatic suspension systems for enhanced performance in mining dump trucks  
  by Handui Feng, Kunliang Yi, Xiumei Liu, Beibei Li, Shen Liu 
  Abstract: This study focuses on optimizing the hydro-pneumatic suspension system of mining dump trucks to enhance dynamic performance. A mechanicalhydraulic integrated half-vehicle model of XDR80T truck was developed in AMESim. Field experiments collected real-time displacement, acceleration, and oil pressure data under loaded and unloaded driving conditions. Theroad spectrum was calculated using the frequency-domain integration method to provide input signals for numerical simulations. Results show a maximum deviation of 6% of Root Mean Square (RMS) values of acceleration signals between simulation and experimental data, confirming the model’s reliability. Further analysis using a random pavement model and factorial analysis explored the impact of basic dimensions, initial air pressure, and inflation volume on suspension performance. Optimization via genetic algorithms minimized both vertical and lateral accelerations of the vehicle body, achieving RMS reductions of 17.65% and 13.83% under heavy-load conditions, and 14.68% and 5.62% when unloaded, highlighting significant suspension performance improvements.
  Keywords: mining dump truck; hydro-pneumatic suspension; sensitivity analysis; structure optimization; genetic algorithm; ride comfort.
  
  
• FraSNA-Net: traffic aware content caching vehicular social networks using fractional shuffle neuron attention network  
  by Vedha Vinodha Doraiswamy, Malathy Subramanium 
  Abstract: This manuscript presents an innovative Deep Learning (DL)-based traffic aware content caching system for Vehicular Social Networks (VSN) called Fractional Shuffle Neuron Attention Network (FraSNA-Net). Primarily, the system model of traffic aware content caching is considered. The content provider-aware routing for the vehicle is performed using hybrid FraSNA-Net by considering multi-objective fitness formulated by parameters including amount of content providers in a path, distance, and traffic. In this case, FraSNA-Net is developed by amalgamating ShuffleNet and Neuron Attention Stage-by-Stage Net (NASNet). Later, an efficient content recommendation is performed by content catching at a content provider. Finally, the content provider distribution is performed using FraSNA-Net, where the weights of FraSNA-Net are effectively trained using Taylor Flamingo Shark Optimization (TFSO). The efficacy of FraSNA-Net is examined by using metrics such as delivery rate, delivery delay, and computational cost, and it gained values of 87.805, 0.640 sec, and 1.205.
  Keywords: content catching; VSN; vehicular social networks; IoV; internet of vehicles; deep learning; vehicle-to-infrastructure communication.
  DOI: 10.1504/IJHVS.2026.10078757
   
  
• Modelling and manufacturing of a polyurethane non-pneumatic tyre for unmanned ground vehicles  
  by Hossam Ragheb 
  Abstract: Non-pneumatic tyres (NPTs) offer greater durability, puncture resistance, and less service than conventional tyres and are therefore a promising alternative; however, non-pneumatic polyurethane tyres for lightweight ground vehicles and robot platforms have not been fully investigated. A finite element analysis shall be performed to determine the stress and deformation patterns and the behaviour of the load deflection under static and dynamic load conditions. The modelling of tyre material has been tested experimentally by means of uniaxial stress tests with ISO 527-2 standard samples and stress relaxation tests with dynamic mechanical analysis. The results indicate that the developed polyurethane non-pneumatic tyre has the capability of adequate load capacity and structural strength while maintaining an acceptable quality of ride, and these findings contribute to the evolution of non-pneumatic tyre technologies and their application in autonomous vehicles, unmanned aerial vehicles, and specialised mobility platforms where puncture reliability and resistance to puncture are critical.
  Keywords: non-pneumatic; polyurethane; finite element analysis; unmanned ground vehicle; robots.
  DOI: 10.1504/IJHVS.2026.10079038
   
  
• Analysis of comfort performance of a multi-layered seat cushion via finite element methods  
  by Darshan Dorugade, Subhash Rakheja, Paul-Émile Boileau, Yumeng Yao, Zhou Zheng 
  Abstract: The comfort performance of a polyurethane foam (PUF) seat cushions is related to highly nonlinear visco-elastic properties of PUF, which are further affected by seat load and deformation in the highly complex manners. This study presents a design concept of a multi-layered PUF seat cushion. A finite element (FE) model of the proposed concept is formulated and analysed to determine body-seat contact pressure distribution. The PUF material model was developed on the basis of the laboratory-measured force-deflection properties of the seat cushion. The seat cushion model developed in the LS-DYNA platform was validated using measured data acquired with two human subjects. The model validation was established 50th percentile dummy with seated mass similar to that of subjects used in the experiment. The model was subsequently refined to incorporate multiple PUF layers in an attempt to reduce the peak contact pressure, and thereby, improve comfort performance. The results showed reduction in contact pressure, especially in ischium region while increasing the contact area. These proposed model can serve as a valuable design guidance for commercial vehicles seats.
  Keywords: seat pressure distribution; multi-layer seat cushion; FE seat cushion model.
  DOI: 10.1504/IJHVS.2025.10079159
   
  
• Stanley-RL: a trajectory tracking control for unmanned ground vehicles  
  by Islam Hassan, Tamer Attia, A.M. Sharaf, Hossam Ragheb 
  Abstract: In this paper, the Stanley method first drives the Unmanned Ground Vehicle (UGV) predetermined trajectory and then the Deep Reinforcement Learning (DRL) corrects the control inputs so that the trajectory is followed with high precision. The proposed control framework uses two neural networks. (NNs): the critic network, which acts as a controller for generating the UGV's actions, and a target network that assesses future rewards. The new approach is tested by means of a UGV traveling on an elliptical track with 420 m in length at speeds between 1 m/s to 5 m/s. The results demonstrate that the proposed control framework improves the performance of the UGV by improving the accuracy of the trajectory tracking. A comparison between using only the Stanley method and the combined approach involving Double Deep Q-Network (DDQN) reveals that the Stanley-based DDQN constitutes a robust framework to achieve accurate and adaptive trajectory tracking in UGV.
  Keywords: Stanley-RL; UGV; unmanned ground vehicle; DRL; deep reinforcement learning; trajectory tracking; DDQN; double deep Q-network.
  DOI: 10.1504/IJHVS.2026.10079394
   
  
• Reinforcement learning and neural network PID-based control for semi-active seat suspension of mountainous-terrain tractors  
  by GuiYong Wang, ZhiGang Wang, ZhiWei Li, RaoQiang Li, Chen Cai 
  Abstract: To enhance ride comfort for mountain tractor operators and improve the energy efficiency of the seat suspension system, a co-simulation model integrating a magnetorheological damper (MRD) within a semi-active seat suspension was developed in Matlab/Simulink. This model incorporates a novel hybrid frequency-event-driven hilly-terrain excitation profile, a five-degree-of-freedom coupled 'tractor-cab-seat' dynamic model, and an MRD characterized by the Bouc-Wen model. A multi-objective optimized RBF-PID controller enhanced with reinforcement learning (RL-RBF-PID) was designed to synergistically suppress vibration and manage energy dissipation. The controller features a dual-time-scale architecture: a Q-learning upper layer for global PID parameter optimization using a composite reward function, and an RBF neural network lower layer for real-time compensation. Comparative experiments under three distinct operational conditions demonstrate the superiority of the RL-RBF-PID strategy. Under the representative mountainous-hilly road excitation, it reduces the RMS values of vertical seat acceleration, vertical displacement, and dynamic travel by 51.7%, 64%, and 57.7%, respectively, while achieving a total energy dissipation of 583.06J, confirming its superior performance in vibration attenuation and energy efficiency.
  Keywords: mountainous terrain; active seat suspension system; magnetorheological damper; reinforcement learning; RBF-PID controller; ride comfort.
  DOI: 10.1504/IJHVS.2025.10079773
   
  
• The effect of standard axle parameters on pavement  
  by Hamed Maleki 
  Abstract: The equivalent single axle load factor is one of the key coefficients in pavement design, representing the impact of existing traffic loads on pavement deterioration. This factor can be determined empirically in accordance with the Iranian Asphalt Pavement Design code, and designers often rely on their experience when selecting its value. In this paper, sensitivity analysis is used to examine the parameters affecting the standard axles. These parameters include the terminal serviceability index, pavement design life, and the pavement's structural number. As the terminal serviceability index increases while other factors remain constant, the number of equivalent axles decreases. As the pavement design life increases while other factors are held constant, the number of equivalent axles increases. When the structural number of the pavement increases and other factors remain fixed, a sinusoidal pattern is observed: The number of equivalent axles decreases up to SN=4, and beyond SN=4, it increases.
  Keywords: AASHTO; tandem axle; tridem axle; structural number; design life; terminal serviceability index.
  DOI: 10.1504/IJHVS.2026.10080188
   
  
• Reconfigurable multi-unit, multi-axle, multi-articulation heavy vehicle (RUAAHV) parameter estimation  
  by Tokologo Komana, Schalk Els, Herman A. Hamersma 
  Abstract: Performance-Based Standards (PBS) is an alternative approach of regulating heavy vehicles focusing on safety performance, rather than prescriptively limiting the vehicle?s dimensions or mass. Commonly, PBS assessments are through simulation requiring accurate vehicle parameters that are not readily available. This study presents a reconfigurable methodology to estimate heavy vehicle states (i.e. slip angle and hitch angle) and parameters (i.e. cornering stiffness and yaw moment of inertia). The state estimator uses a kinematic Kalman Filter (KF) and the parameter estimator employs a dynamic model to estimate the parameters of a tractor and semi-trailer from experimental data. The estimated parameters are used to generate an Adams View model of the experimental vehicle, which is validated against a different set of experimental data. Results show that the proposed Reconfigurable multi-Unit, multi-Axle, multi-Articulation Heavy Vehicle (RUAAHV) parameter estimator will aid in estimating important heavy vehicle parameters necessary to perform PBS assessments accurately.
  Keywords: parameter estimation; state estimation; PBS; performance-based standards; heavy vehicle dynamics; cornering stiffness; yaw moment of inertia.
  DOI: 10.1504/IJHVS.2026.10080347
   
  
• Autonomous vehicle navigation for dynamic environments: an AI-powered framework with PDQN and multi-objective hybrid optimisation  
  by Jagadeeshwar Podishetti, Sridhar Chintala, Srihari Gude, Rajendar Kalvala 
  Abstract: This research presents an AI-powered navigation system for AVs designed to operate competently in dynamic environments. In proposed navigation system, technologies are integrated including Deep Learning (DL) for perception, task scheduling and path planning via Multi-Objective Optimization (MOO), adaptive navigation in real-time, and dynamic conditions using Deep Reinforcement Learning (DRL). After vehicles initialization, an advanced Transformer-Based Obstacle Detection System (TODS) combined with Vision Transformers (ViTs) enhances real-time obstacle detection and collision avoidance. Then, a Prioritized Double Q-Network (PDQN) is employed to maintain strong and adaptable approach to path planning. For optimizing objectives such as path length, energy consumption, and overall mission efficiency, a Multi-Objective Genetic-Particle Optimizer (MOGPO) is proposed. Also, system incorporates an intelligent recharging strategy to ensure vehicle autonomously navigate to charging stations when necessary. This study strives to contribute reliable and scalable solution to AVs by augmenting their capacity to function effectively under dynamic and real-life situations.
  Keywords: autonomous vehicles; deep learning; MOO; multi-objective optimisation; DQN; deep Q network; transformers.
  DOI: 10.1504/IJHVS.2026.10080348
   
  
• Rollover prevention of heavy vehicles using CMG-based neural LQR gain scheduling under time-varying operating conditions
  by Faruk Ünker 
  Abstract: Heavy vehicles with high centres of gravity are highly susceptible to rollover under banked road conditions, particularly when system parameters vary dynamically. This study proposes a control moment gyroscope (CMG)- based stabilisation framework enhanced by neural network-driven gain scheduling. Vehicle dynamics are modelled via Lagrangian formulation and linearised for control design. Three strategies are compared: proportional control, optimised linear quadratic regulation (LQR), and a neural network-based LQR (NNLQR) using a parallel independent network (PIN) architecture. While LQR achieves stable roll regulation under constant rotor speeds, it fails under time-varying conditions. The proposed NNLQR maintains stable closed-loop behaviour across [100, 3000] rad/s, with relative gain errors below 0.4%. Comparative simulations demonstrate a computational speed-up of four orders of magnitude over real-time Riccati solutions. The approach provides a practical, embedding-compatible solution for adaptive vehicle stabilisation under dynamic operating conditions.
  Keywords: control moment gyroscope; vehicle rollover prevention; LQR optimization; neural network control; gain scheduling; vehicle dynamics.
  
  
• Study on the influence of planet carrier support on load-sharing and vibration of planetary gears  
  by Wang Yunyang, Ping Zhou, Lei Meng 
  Abstract: The floating design of the planetary carrier is an effective approach to reduce its eccentricity error, primarily achieved by controlling bearing clearance or support stiffness. However, the effects of these two methods on the load-sharing characteristics and vibration performance of planetary gears have not been thoroughly studied. To address this issue, a dynamic model of a straight-tooth planetary gear transmission system was constructed using dynamic simulation software. Different planetary carrier support properties impacts on mean load coefficient and vibration were analysed and experimentally validated. Results indicate that both the load-sharing coefficient and vibration intensity peak at the resonance frequencies of the planetary gears. Controlling planet carrier support stiffness improves load distribution and vibration more effectively than adjusting bearing clearance. The floating design improves the load-sharing and vibration characteristics of planetary gears at low and medium speeds, however, at high speeds, the amount of floating should be minimised.
  Keywords: vibration； planetary carrier； planetary gear； support characteristics； eccentric mass； load-sharing.
  DOI: 10.1504/IJHVS.2025.10080587