Abstract: Battery Thermal Management Systems (BTMS) are critical for maintaining optimal operating temperatures (20-40°C) in lithium-ion batteries, particularly for electric vehicles (EVs) and grid-scale energy storage [1,2]. Phase Change Materials (PCMs) have emerged as a transformative solution, leveraging latent heat absorption/release during phase transitions to provide passive thermal regulation [3]. This review systematically evaluates inorganic (salt hydrates), organic (paraffins, fatty acids), and composite PCMs, analyzing their thermophysical properties, performance characteristics, and implementation challenges in BTMS applications [4,5]. Key findings reveal that advanced composite PCMs with thermal conductivity enhancers (graphene, metal foams) can achieve 3-5× improvement in heat dissipation while maintaining >90% of base latent heat capacity [6,7]. The paper concludes with actionable recommendations for next-generation PCM development and integration strategies.

Abstract: This paper presents a practical industrial hybrid control architecture that augments the widely deployed 49-rule Mamdani fuzzy supervisory PID controller with a lightweight online meta-tuner based on Soft Actor-Critic (SAC) reinforcement learning. While the inner 1 kHz fuzzy-PID loop remains fully deterministic and identical to the industrial baseline, a separate 10 Hz SAC agent autonomously adapts the three output scaling factors (α_Kp, α_Ki, α_Kd ∈ [0.5, 2.5]) of the fuzzy layer using an ONNX Runtime inference engine. The complete controller is implemented and experimentally validated on a real Siemens S7-1214C PLC (6ES7214-1AG40-0XB0) in a hardware-in-the-loop setup with a high-fidelity 5-DoF manipulator model incorporating measured friction, backlash, sensor noise, and payload variation (0–2.5 kg). Across four demanding scenarios (sinusoidal tracking, sudden payload jumps, sustained disturbances up to 0.76 Nm, and high-speed motions), the proposed method consistently achieves 46–52 % lower RMSE and 28–30 % reduced control energy compared to the fixed-scaling industrial baseline, while preserving strict real-time constraints (inner loop cycle time 0.68–0.89 ms, SAC inference < 0.6 ms). The full PLC program (SCL/FBD), HIL environment, and trained policies will be released open-source upon acceptance (DOI to be provided during revision).The full PLC program, HIL environment, and trained SAC policies will be released open-source as a preprint supplement.

Abstract: Infrared traffic object detection faces challenges such as low resolution, weak thermal 2 contrast, and inefficiency in detecting small objects. To address these issues, this paper 3 proposes RES-YOLO, an enhanced YOLOv8n-based architecture. It incorporates Receptive 4 Field Adaptive Convolution for improved multi-scale perception, Efficient Multi-scale 5 Attention for better feature representation, and the Scylla-IoU loss for more accurate 6 and faster bounding box regression. Additionally, a pseudo-color infrared dataset is 7 constructed to enrich texture and contrast information beyond conventional white-hot 8 images. Experiments on both the FLIR public dataset and a self-built dataset show RES- 9 YOLO improves accuracy by 4.9% and 5.5% over the baseline while maintaining real-time 10 performance. These results highlight the method’s effectiveness in integrating lightweight 11 deep learning and dataset enhancement for robust perception in intelligent vehicle systems, 12 supporting AI-driven autonomous driving and driver assistance applications.

Abstract: Development of AI-driven automated driving functions requires vast amounts of diverse, high-quality data to ensure road safety and reliability. However, manual collection of real-world data and creation of 3D environments is costly, time-consuming, and hard to scale. Most automatic environment generation methods still rely heavily on manual effort, and only a few are tailored for Advanced Driver Assistance Systems (ADAS) and Automated Driving (AD) training and validation. We propose an automated generative framework that learns real-world features to reconstruct realistic 3D environments from a road definition and two simple parameters for country and area type. Environment generation is structured into three modules - map-based data generation, semantic city generation, and final detailing. The overall framework is validated by training a perception network on a mixed set of real and synthetic data, validating it solely on real data, and comparing performance to assess the practical value of the environments we generated. By constructing a Pareto front over combinations of training set sizes and real-to-synthetic data ratios, we show that our synthetic data can replace up to 90% of real data without significant quality degradation. Our results demonstrate how multi-layered environment generation frameworks enable flexible and scalable data generation for perception tasks while incorporating ground-truth 3D environment data. This reduces reliance on costly field data and supports automated rapid scenario exploration for finding safety-critical edge cases.

Abstract: The shift toward Industry 5.0 places human-centred and digitally integrated metrology at the core of modern manufacturing, particularly in the automotive sector, where portable Laser Line Triangulation (LLT) systems must combine accuracy with operator usability. This study addresses the challenge of operator-induced variability by evalu-ating how algorithmic strategies and mechanical support features jointly influence the performance of a portable LLT device derived from the G3F sensor. A comprehensive Measurement System Analysis was performed to compare three feature-extraction al-gorithms—GC, FIR, and Steger—and to assess the effect of a masking device designed to improve mechanical alignment during manual measurements. The results highlight distinct algorithm-dependent behaviours in terms of repeatability, reproducibility, and computational efficiency. More sophisticated algorithms demonstrate improved sensi-tivity and feature localisation under controlled conditions, whereas simpler gradi-ent-based strategies provide more stable performance and shorter processing times when measurement conditions deviate from the ideal. These differences indicate a trade-off between algorithmic complexity and operational robustness that is particu-larly relevant for portable, operator-assisted metrology. The presence of mechanical alignment aids was found to contribute to improved measurement consistency across all algorithms. Overall, the findings highlight the need for an integrated co-design of algorithms, calibration procedures, and ergonomic aids to enhance repeatability and support operator-friendly LLT systems aligned with Industry 5.0 principles.

Abstract: With the rapid development of the automotive industry, autonomous driving has attracted growing research interest, among which path planning and trajectory tracking play a central role. To better understand the evolution, current status, and future directions of this field, this study conducts a comprehensive bibliometric analysis combined with latent Dirichlet allocation (LDA) topic modeling on publications related to autonomous vehicle path planning and trajectory tracking indexed in the Web of Science database. Multiple dimensions are examined, including publication trends, highly cited authors, leading institutions, research domains, and keyword co-occurrence patterns. The results reveal a sustained growth in research output, with trajectory planning, path optimization, trajectory tracking, and model predictive control emerging as dominant topics, alongside a notable rise in learning-based approaches. In particular, reinforcement learning and deep reinforcement learning have become increasingly prominent in complex decision-making and tracking control scenarios. The analysis further identifies core contributors and institutions, highlighting the leading roles of China and the United States in this research area. Overall, the findings provide a systematic overview of the knowledge structure and evolving research trends, offering valuable insights into key opportunities and challenges and supporting future research toward safer and more intelligent autonomous driving systems.

Abstract: We present an adaptive governance engine for blockchain-enabled Vehicular Ad Hoc Networks (VANETs) that manages latency-energy-coherence trade-offs under rapid topology changes3. We introduce (i) an Ideal Information Cycle as an operational abstraction of information injection and validation and (ii) a modular VANET Engine implemented as a real-time control loop in NS-3.354. The Engine monitors normalized Shannon entropies—informational entropy $S$ over active transactions and spatial entropy $H_{spatial}$ over occupancy bins (both on [0, 1])—and adapts the consensus mode (PoW versus signature/quorum-based modes such as PoS/FBA) together with key rigor parameters via calibrated policy maps5. Governance is cast as a constrained operational objective that trades per-block resource expenditure (radio + cryptography) against a Quality-of-Information (QoI) proxy derived from delay/error tiers, subject to timeliness and ledger-coherence pressure6. Cryptographic cost is explicitly traceable through counted operations, $E_{crypto} = e_h n_{hash} + e_{sig} n_{sig}$, and coherence is tracked via an LCA/LCP-normalized ledger-divergence metric7. We evaluate the framework under urban/highway mobility, scheduled partitions, and bounded adversarial stressors (Sybil identities, Byzantine proposers), using 600 s runs with 30 matched random seeds per configuration and 95% bootstrap confidence intervals8. In high-disorder regimes ($S \ge 0.8$), the Engine reduces total per-block energy (radio + cryptography) by more than 90% relative to a fixed-parameter PoW baseline tuned to the same latency target9. A consensus-first triggering policy further lowers agreement latency and increases throughput compared with broadcast-first baselines10. Under high mobility ($v=30$ m/s) in the emphasized urban setting, the Engine bounds orphaning ($\le 10\%$), keeps finality within sub-150 ms ranges, and reduces average ledger divergence below 0.07 at high spatial disorder11. Scope and security envelope: the main evaluation is limited to $N \le 100$ vehicles under full PHY/MAC fidelity12. PoW targets are intentionally latency-feasible and are not intended to provide cryptocurrency-grade majority-hash security; operational security and mode-transition considerations are discussed explicitly in Section 413.

Abstract: Background: The automotive industry is undergoing a deep transformation driven by 9 the global green transition. This change follows divergent trajectories in developed and 10 emerging markets due to differences in regulation, infrastructure, and economic con- 11 straints. The research methodology is adapted to incorporate different factors of influence 12 and contraints. The research applies a structured Failure Mode and Effects Analysis 13 (FMEA) based on IEC 60812:2018 and AIAG & VDA (2019), and integrates the Analytic 14 Hierarchy Process (AHP) to prioritize corrective measures. Concepts from adaptive risk 15 management, informed by expert consensus and literature-backed data, are also used to 16 interpret dynamic behavior of supply chains and market volatility. The comparative anal- 17 ysis successfully highlights the systematic RPN Divergence between market types, reveal- 18 ing critical differences in failure mode profiles, risk priorities, and capacity to adopt miti- 19 gation strategies. The hybrid FMEA-AHP approach reduces subjectivity and provides 20 transparent prioritization tailored to market maturity. The integrated methodology sup- 21 ports decision-making in electrification programs and offers a robust framework for 22 benchmarking complex transition processes across regions.

Abstract: The objective of the article is to present investigates the feasibility of driver-state assessment in a real automotive environment using a mobile long-wave infrared (LWIR) thermal camera. Unlike visible-spectrum systems, thermal imaging provides illumination-invariant and temperature-dependent information that is particularly advantageous inside a vehicle, where lighting conditions vary substantially. A handheld microbolometer (UTi260M) was used to record thermal video of a driver during prolonged, monotonous driving with a stabilized cabin temperature. Pixel-wise temperature reconstruction, spatial noise estimation, uniformity analysis, and NETD approximation were applied to evaluate thermal image quality and to quantify thermophysiological changes associated with drowsiness. The thermal recordings revealed characteristic pre-sleep markers, including head droop, reduced neuromuscular correction, elevated and spatially uniform facial temperature, and diminished thermal variability. These patterns correspond to known physiological responses to fatigue, reduced sympathetic activation, and warm cabin exposure. The analysis demonstrates that mobile thermal imaging can reliably capture early indicators of declining vigilance and can support the development of non-contact driver-monitoring systems. The findings further suggest that integrating temperature-driven alerting thresholds into mobile applications may provide an additional preventive mechanism against drowsiness-related accidents.

Abstract: Vehicle exhaust gases remain one of the key sources of atmospheric air pollution and pose a serious threat to ecosystems and public health. This study presents an experimental investigation into reducing the toxicity of gasoline internal combustion engine exhaust using ultrasonic waves and infrared (IR) laser exposure. An original hybrid system integrating an ultrasonic emitter and an IR laser module was developed. Four operating modes were examined: no treatment, ultrasound only, laser only, and combined ultrasound–laser treatment. The concentrations of CH, CO, CO2, and O2, as well as exhaust gas temperature, were measured at idle and under operating engine speeds. The experimental results show that ultrasound provides a substantial reduction in CO concentration (up to 40%), while IR laser exposure effectively decreases unburned hydrocarbons CH (by 35–40%). The combined treatment produces a synergistic effect, reducing CH and CO by 38% and 43%, respectively, while increasing the CO₂ fraction and decreasing O₂ content, indicating more complete post-oxidation of combustion products. The underlying physical mechanisms responsible for the purification were identified as acoustic coagulation of particulates, oxidation, and photodissociation of harmful molecules. The findings support the hypothesis that combined ultrasonic and laser treatment can enhance real-time exhaust gas purification efficiency. It is demonstrated that physical treatment of the gas phase not only lowers the persistence of by-products but also promotes more complete oxidation processes within the flow.

Abstract: This paper introduces a novel Magnetorheological (MR) damper integrating a ball-screw mechanism (SMRB damper), designed to unify translational and rotational motion for enhanced automotive suspension performance. While shear-mode rotary MR dampers offer excellent responsiveness and stability, prior designs face persistent issues such as high off-state torque, structural complexity, or limited damping force. The proposed damper aims to overcome these limitations. Its design and operating principle are presented, followed by the development of a mathematical model based on the Bingham-plastic formulation and finite element analysis. To maximize damping capability, the key structural parameters are optimized using an Adaptive Particle Swarm Optimization (APSO) algorithm. Finally, a prototype is fabricated based on the optimized results, and experimental tests validate its performance against simulation predictions, demonstrating its improved potential for vibration control applications.

Abstract: Biodiesel fuel produced through transesterification is mainly used in blends with conventional diesel fuel. The analysis of combustion process parameters for each specific biodiesel fuel represents the basis for a rational approach to the utilization of available motor fuel quantities. In this study, the heat release rate and cumulative heat release during the combustion of conventional diesel fuel and blends of biodiesel fuel made from waste grape seed oil and conventional diesel fuel were analyzed. The tests were conducted on a single-cylinder, air-cooled diesel engine with direct fuel injection. The combustion of conventional diesel fuel, a blend containing 7% of biodiesel by volume (B7), and a blend containing 14% of biodiesel by volume (B14) was examined. Using blends, especially those with a higher biodiesel content (B14), results in a higher maximum heat release rate compared to conventional diesel fuel, which can have negative implications in terms of mechanical stresses and engine noise. However, the higher combustion rate of the B14 blend, particularly during the combustion of the first 50% of the fuel mass per cycle, can have a positive impact on the fuel economy of the working cycle and the engine as a whole.

Abstract: Motor position sensors are critical parts for traction motors control in electrified automotive powertrain. As motors are getting more compact due to the advance of technology the packaging space for motor position sensors is becoming increasingly restricted. This study presents a two-layer (2L) printed circuit board (PCB) routing strategy for inductive motor position sensors with limited area. A prototype was fabricated and tested on a test bench using a comprehensive design of experiments that contains 625 combinations of X- and Y-offsets, tilt angle, and airgap at various levels (±0.5 mm in X/Y, ±0.5° tilt, 1.9–3.1 mm airgap). Across the tolerance box, the accuracy under all test cases remained within ±1 electrical degree. The accuracy analysis through Fourier series on circle shows that the DC offset and magnitudes mismatch of the 3 Rx signals are the dominant error contributors due to the routing modification. An Extreme Gradient Boosting (XGBoost) model was trained and validated with R² = 0.9951 on the sensor data. The SHapley Additive exPlanations (SHAP) analysis identified tilt and Y-offset as dominant contributors to accuracy degradation. The model revealed a mild Y-axis asymmetry introduced by routing modifications. The SHAP results show that machine learning workflow provides a general, quantitative framework for analyzing inductive sensor layouts and installation tolerances.

Abstract: Fuel consumption in heavy-duty off-road machinery depends on a wide range of in-teracting factors related to the operating environment, the technical characteristics and condition of the machine, and the behaviour, experience and state of the operator. Existing studies typically address only fragments of this relationship, focusing on vehicle parame-ters, selected environmental factors or individual aspects of driving style. The method proposed in this work provides a general and transferable framework for assessing fuel consumption in any type of machine or vehicle. The Integrated Fuel Consumption As-sessment Model (IFCAM) combines environmental, vehicle and human domains into a coherent structured formula that can be used across different operational contexts. The model was developed using continuous short-term measurements and long-term opera-tional data collected during real industrial work. Its universal structure makes it applica-ble not only to mining equipment, but also to construction machinery and transport vehi-cles, as well as conventional passenger cars, where it offers a systematic procedure for es-timating fuel demand under variable operating conditions. The results demonstrate that integrating multi-domain data improves predictive accuracy and opens new possibilities for analysing operator influence and overall energy efficiency.

Abstract: Transmission error (TE) is one of the most important sources of gear noise and vibration. Manufacturing tolerances and assembly shifts introduce deviations in tooth geometry that produce periodic mesh disturbances; these disturbances excite the drivetrain and radiate as airborne sound. While many studies have modelled individual tolerance effects in deterministic simulations, few have evaluated the combined influence of realistic tolerance distributions using open Monte–Carlo data. This work analyses the public Gear Statistical tolerance analysis dataset (≈40 k samples) to answer the following questions: (Q1) How accurately can the kinematic transmission error be predicted from measured tolerance and process‑shift data? (Q2) Which individual tolerances and their interactions have the greatest impact on TE? (Q3) Can a TE‑derived proxy quantify noise‑critical excitation without an acoustic model? (Q4) What tolerance combinations minimise TE and noise while respecting manufacturing cost? (Q5) Is there a quantifiable trade‑off between tolerance tightening cost and noise reduction? To address these questions we formulate hypotheses: H1 Non‑linear machine‑learning models achieve high predictive accuracy for TE (R² > 0.85) compared with linear baselines. H2 A small subset of tolerances—profile and lead errors—dominate the TE variance and interact non‑linearly. H3 A relative noise proxy (RNP) derived from TE faithfully ranks noise‑critical excitation across tolerance combinations. H4 A Pareto front exists in the cost–noise plane, enabling cost‑effective tolerance optimisation. H5 Targeted adjustment of the top two tolerances yields larger noise‑reduction per cost than uniform tightening. The following sections describe the data, the modelling framework and the results that support these hypotheses.

Abstract: Improving the ride comfort of commercial vehicles is crucial for driver health and operational safety.This study focuses on optimizing the parameters of a cab suspension system to improve its vibration isolation performance. Initially, nonlinear fitting was applied to experimental data characterizing air spring stiffness and damping, which informed the development of a multi-body rigid-flexible coupled dynamic model of the suspension system; its dynamic characteristics were subsequently validated through modal analysis. Road excitation data, filtered through the chassis suspension, were collected during vehicle testing, and displacement excitations for ride comfort simulation were reconstructed using virtual iteration technology. Thereafter, an integrated ISIGHT platform, combining ADAMS and MATLAB, was employed to systematically optimize suspension parameters and key bushing stiffness via a multi-island genetic algorithm. The optimization results demonstrated significant performance improvements: on General roads, the overall weighted root-mean-square acceleration was markedly reduced with enhanced isolation efficiency; on Belgian pave roads, resonance in the cab's X-axis direction was effectively suppressed; and on Cobblestone roads, the pitch angle was successfully constrained within the design limit. This research provides an effective parameter matching methodology for performance optimization of cab suspension systems.

Abstract: Radiated noise from gearbox housings is a significant contributor to the noise, vibration, and harshness (NVH) performance of automotive drivetrains. Small deviations in wall thickness caused by standard manufacturing tolerances can influence structural stiffness, shift modal frequencies, and alter acoustic radiation levels, particularly near resonance. This study numerically examines the effect of ±10% wall thickness variation using a simplified multi-mode structural–acoustic model. Modal frequencies were scaled proportionally to thickness changes, and frequency response functions were calculated for nominal and tolerance-limit geometries. Statistical variability was assessed through Monte Carlo simulations with wall thickness values sampled from a truncated normal distribution within the tolerance range. Results show that a ±10% variation can produce modal frequency shifts accompanied by resonance amplitude differences of up to 20.9 dB in the 950–1050 Hz range, with an average change of 3.34 dB between extreme cases. Monte Carlo analysis indicated mean changes of 0.54 dB in the main resonance band, with some cases reaching 2.54 dB. The findings demonstrate that even within standard production limits, wall thickness tolerances can measurably affect gearbox NVH behavior. Considering such variability in early-stage design can help reduce unit-to-unit noise differences and improve acoustic robustness.

Abstract: Accurate detection and localization of traffic objects are essential for autonomous driving tasks such as path planning. While semantic segmentation is able to provide pixel-level classification, existing networks often fail under challenging conditions like nighttime or rain. In this paper, we introduce a new training framework that combines unsupervised domain adaptation with high dynamic range imaging. The proposed network uses labeled daytime images along with unlabeled nighttime HDR images. By utilizing the fine details typically lost in conventional SDR images due to dynamic range compression, and incorporating the UDA training strategy, the framework effectively trains a model which is capable of semantic segmentation across the adverse weather conditions. Experiments conducted on four datasets have demonstrated substantial improvements in inference performance under nighttime and rainy scenarios. The accuracy for daytime images is also enhanced through expanded training diversity. Source code is available at https://github.com/ZackChen1140/RMSeg-HDR.

Abstract: Income tax fraud is a serious challenge to revenue authorities, which involves large amounts of money losses and erodes public trust. The conventional methods of detection through manual audits and rule-based systems tend to be slow and ineffective for large volumes of data. This study suggests an AI - and ML-based framework for the detection of tax fraud through fraudulent returns by using both supervised and unsupervised learning algorithms. Major models adopted are Random Forest, XG-Boost, and Isolation Forest, selected based on their performance in classification and outlier detection. Feature engineering is directed toward significant features like income patterns, deductions, exemptions, and past filing behavior to detect abnormal patterns that signify fraud. Experimental outcomes reveal that Random Forest performed best with an accuracy of 96%, followed by XG-Boost with an accuracy of 95%, and Isolation Forest with an accuracy of 80%, showing the best performance of tree-based ensemble models for the task. The system provides a risk score for every tax return, allowing authorities to rank audits and reduce false positives. These results show that machine learning models far surpass conventional methods, delivering a scalable and automated solution for effective fraud detection. The research provides a realistic basis for incorporating AI-based strategies in financial fraud management, leading to increased compliance, minimized revenue leakage, and better decision-making by tax authorities.

Abstract: Lightweight gearbox housings often raise NVH risk, yet full finite-element evaluations are too slow for early design screening. This study tests whether a few frequency-band descriptors of radiated sound are enough to classify housing stiffness. Using an open dataset of electric-vehicle gearbox spectra for three rib-configurations—flexible, intermediate and rigid—we averaged sound-pressure levels in five 1 kHz bands. Principal-component analysis separated the twelve samples into three non-overlapping groups, confirmed by k-means clustering (adjusted Rand index = 1.00). The random-forest model achieved 75 % classification accuracy on the present 12-sample data set (leave-one-out evaluation). Owing to the small sample size this figure should be regarded as explorative, and a larger validation study is required to confirm generalizability; permutation analysis confirmed the 3–4 kHz and 2–3 kHz bands as most important for classification. In contrast, total integrated spectral energy showed no significant group difference (p = 0.81). The results These findings suggest that mid-frequency band energy may encode structural-stiffness differences, although validation on larger datasets is necessary. The workflow—load spectra, compute five band means, classify—offers a rapid, interpretable tool for NVH-aware lightweight design.