International Journal of Vehicle Performance (9 papers in press)

Regular Issues

• Road friction coefficient estimation fusing vehicle dynamics and GA-BP neural network  
  by Bin Huang, Zhuang Wu, Wenbin Yu, Zeyang Zhang, Ansheng Yang 
  Abstract: Road friction coefficient is critical to vehicle traction, braking distance, handling stability and active safety system performance, but existing estimation methods have notable flaws: onboard sensors are easily disturbed, vehicle dynamics models depend on idealised assumptions, and data-driven methods overlook variable correlations. This paper builds a 3-DOF vehicle dynamics model and Dugoff tyre model to derive the nonlinear relationship between road friction coefficient and multiple vehicle state variables for neural network inputs. It further optimises traditional BP neural networks with genetic algorithms to construct a GA-BP estimator, overcoming local optima issues. Validated via 441-condition, 4.41-million-sample CarSim-MATLAB/Simulink joint simulations and serpentine tests, GA-BP converges faster than BP, cutting MAE by 35.0% and RMSE by 38.6% on average, delivering a reliable real time friction coefficient estimation solution.
  Keywords: road friction coefficient; vehicle dynamics; neural network; generalisation ability; genetic algorithm; GA.
  
  
• Moroccan licence plate detection and recognition using a YOLOv10s detector and a dual-CNN module  
  by Ouafae Abarkan, Walid Jebrane, Ihssane Bouasria, Nabil El Akchioui 
  Abstract: The recognition of licence plates using artificial intelligence has been a recurring research topic. It plays a significant and important role in various fields and offers road safety benefits, particularly in intelligent traffic monitoring systems, parking management, toll management, access control to public or private locations, and identification of vehicles in violation. In this paper, we implement a practical application for the detection and recognition of Moroccan licence plates from images captured from the front or rear of vehicles, highlighting the challenges related to their format, font, and diversity. For the detection plates, we used the YOLOv10s model, a recent architecture that enables accurate and real-time plate localisation. For this purpose, we developed and trained two convolutional neural network (CNN) models: the first dedicated to the recognition of Arabic letters and the second to the recognition of numbers. These models work complementarily to ensure more effective licence plate recognition.
  Keywords: intelligent systems; licence plates; convolutional neural networks; CNNs; YOLOv10s; deep learning.
  DOI: 10.1504/IJVP.2026.10078949
   
  
• OEM-level experimental assessment of worn-tyre braking performance under UNECE R117 conditions  
  by Barış Uluçay, Berkay Cedetaş 
  Abstract: Worn tyres increasingly pose both safety and environmental risks, as addressed by UNECE Regulation No. 117 and the forthcoming EURO 7 requirements. This study quantitatively evaluates the effect of tyre wear on dry braking performance of passenger car tyres under controlled OEM test conditions. Braking tests were conducted on a Fiat Egea 1.6 L at 80 km/h ( 0.9) under controlled asphalt conditions, comparing new (7.8 mm tread depth) and worn (1.6 mm tread depth) tyres using a VBOX 2 SL data logger. Results showed an 8% increase in braking distance for worn tyres (p = 0.02), consistent with performance losses reported in the literature, particularly in grip. Findings confirm that tyre wear significantly affects braking dynamics, even under dry conditions, and highlight the importance of integrating worntyre validation into OEM development and regulatory approval processes. Moreover, the relationship between tread depth and braking performance is nonlinear, with rapid deterioration in braking, wet grip, and wet braking as tyres approach the legal wear limit. This work provides a repeatable, OEMlevel methodology that bridges design validation and regulatory assessment, supporting safer and performanceoptimised tyre lifecycle management.
  Keywords: worn tyre performances; braking distances; grip; traction; rolling resistances; environmental impacts; UNECE R117 regulation.
  DOI: 10.1504/IJVP.2026.10079084
   
  
• Enhanced MR suspension systems via optimised cascade PI-PD control  
  by Heba El-Taweel, Sayed Elhussieny, Hassan Metered, Mohamed M. Abd Elhafiz 
  Abstract: This study proposes a cascade PI-PD controller to enhance the performance of an MR-damped semi-active vehicle suspension system. The controller is optimised using a multi-objective salp swarm algorithm (MSSA). The suspension system is modelled as five degrees of freedom (five DOF) half-vehicle, and its equations of motion are simulated in MATLAB/Simulink. Performance is evaluated under double bump and random road excitations, focusing on bounce and pitch dynamics. Key performance indicators, including suspension deflection, body acceleration, and dynamic tyre deflection, are analysed in both time and frequency domains. The MSSA-based PI-PD controller outperforms passive strategies, including conventional passive and zero-voltage MR systems, achieving reductions of up to 26% in body acceleration, 49% in suspension deflection, and 27% in dynamic tyre deflection. Compared with standard PID and skyhook controllers, it provides up to 20% improvements in ride comfort and handling, demonstrating its superior effectiveness across passive and semi-active control approaches.
  Keywords: vehicle suspension; semi-active suspension; magnetorheological damper; MR damper; cascade controller; skyhook control; PID controller; PI-PD controller; multi-objective optimisation; salp swarm algorithm; SSA.
  DOI: 10.1504/IJVP.2026.10079291
   
  
• Research on automatic lane-changing decision of heavy trucks in highways based on Bayesian networks and rough set theory  
  by Libo Mao, Guangqiang Wu 
  Abstract: This paper addresses the lane-changing decision problem of unmanned heavy-duty trucks in highway scenarios and develops a novel algorithm integrating rough set theory and Bayesian networks. Nine conditional attributes related to vehicle speed, relative distance and relative speed are selected, along with three decision categories including left lane change, right lane change and no lane change. The training dataset is built using data from the public HighD dataset and real driving data collected on Donghai Bridge. Attribute values are discretised by the Chimerge algorithm, and redundant attributes are eliminated through rough set reduction. A Bayesian network is then established to learn lane-changing patterns and judge the appropriate lanechanging moment. PreScan-Simulink co-simulations and real-vehicle tests are conducted for comprehensive validation. Results show that the proposed method possesses high decision accuracy and low computational overhead, which fully meets the practical requirements of engineering applications
  Keywords: autonomous heavy truck; rough set theory; Bayesian networks; data discretisation; attribute reduction.
  DOI: 10.1504/IJVP.2026.10079457
   
  
• High-fidelity lap time simulation of Formula One car: multi-circuit validation with real-world telemetry  
  by Fauzan Ahmad Sofyan, M. Nasirudin, Zainal Abidin 
  Abstract: Lap time is the key performance metric in modern motorsport, enabling engineers to accurately predict on-track behaviour and optimise vehicle performance under strict financial regulations and limited testing opportunities. This paper presents the development and multi-circuit validation of a high-fidelity Formula 1 simulation model, designed to act as a reliable digital twin for performance analysis and predictive engineering. Unlike prior studies that focused on go-karts or Formula Student cars, this work validates a full F1 car model against telemetry across five racetracks with diverse characteristics. Results show strong agreement with real-world telemetry data, with deviations kept as low as 0.104 seconds or an error of just 0.13%. Overall, the findings demonstrate that accurate and reproducible lap time prediction of Formula 1 cars can be achieved using accessible simulation tools, providing a robust baseline for subsystem analysis and future vehicle dynamics research.
  Keywords: lap time simulation; vehicle dynamics; simulation validation; digital twin; motorsport.
  DOI: 10.1504/IJVP.2026.10079530
   
  
• A dual energy efficient hybrid control strategy for electric vehicle-IoT driven by in-wheel-motors based on adaptive sliding mode control and medium access control  
  by Mahdi Al-Quran, Yahia Al-smadi, Waleed Nayfeh, Majd Abbad 
  Abstract: In the present work, a dual energy-efficient hybrid control strategy is proposed for the electric vehicle-internet of things (EV-IoT) systems driven by in-wheel motors. The proposed approach integrates the adaptive sliding mode control (ASMC) with the bit mapping buffer status of medium access control (MAC) to optimise the data transmission and energy consumption. A proposed strategy utilises the bit mapping of slide mode control, which dynamically adjusts the operational modes through control signals. The data management is organised into data slots. Each slot contains four sub-slots that accommodate mixed data from the magnetometer, LIDAR and RADAR (MLR) sensors, global navigation satellite system (GNSS), inertial measurement unit (IMU) and barometric altitude sensor (BAS). Each vehicle transmits its bit mapping data to the roadside unit (RSU) which indicates the buffer status of each sensor type are full or empty. This information allows the RSU to selectively activate the radio during the necessary sub-slots. It processes only the relevant data and thus optimises the usage of energy. The proposed approach enhances the overall system performance by ensuring real-time data processing while minimising the energy resources. The proposed approach offers a structured and efficient method for managing sensor data channels. Thus, the present work contributes in the advancement of energy-efficient EV-IoT systems. In the present work, the proposed method is also compared with the existing variants of TDMA protocols. The results prove the superiority of the proposed method.
  Keywords: sliding mode control; SMC; medium access control; MAC; bit-mapping; electric vehicle-internet of things; EV-IoT.
  DOI: 10.1504/IJVP.2026.10079677
   
  
• Design of optimal drive control strategy for extended range electric vehicles  
  by Long Huang 
  Abstract: This paper proposes a drive control strategy based on reference SOC curve and simulated annealing particle swarm optimisation (SA-PSO) to address the problems of high energy consumption, large battery SOC fluctuations, and low drive system efficiency in extended range electric vehicles (EREVs) under complex operating conditions. By combining vehicle dynamics with the equivalent circuit model of the battery, a control model with the goal of minimising energy consumption is constructed. The operating condition reference SOC curve is generated through dynamic programming, and the PSO update process is optimised using SA. The results show that compared with traditional logic threshold and single PSO strategy, this strategy significantly reduces SOC tracking deviation and energy consumption by 9.5%19.8%. Compared to genetic algorithms, it has a 46.2% increase in convergence speed and stronger stability, balancing energy efficiency and system stability.
  Keywords: particle swarm optimisation; PSO; extended-range electric vehicle; drive control strategy; dynamic programming; DP; simulated annealing; SA.
  DOI: 10.1504/IJVP.2026.10079968
   
  
• Designing a terminal sliding mode controller for steer-by-wire system considering vehicle's network delay  
  by Ali Anisi, Moosa Ayati 
  Abstract: This paper presents a fault-tolerant algorithm specifically designed for the steer-by-wire (SbW) system, addressing critical challenges such as random fractional delays and network-induced packet loss issues rarely addressed comprehensively in existing research. The algorithm integrates innovative elements, including a terminal sliding-mode disturbance observer and a discrete-time sliding-mode controller with a time-delay compensator, both optimised for the SbW system using a simplified model. Extensive validation and performance assessment, conducted using the CarSim/MATLAB co-simulator platform, provide unprecedented insights into its capabilities under simulated real-world conditions, marking an advancement in the field.
  Keywords: steer-by-wire; SbW; discrete-time sliding mode controller; networked delay; fault detection; input disturbance.
  DOI: 10.1504/IJVP.2026.10080061