Abstract: To investigate the fracture crack in the rear axle of a certain passenger car during accelerated life testing, a comparative testing analysis was conducted on this car after replacing its new rear axle and on a comparative car. Strain on the rear axle, vibration acceleration, and vehicle speed signal were measured for both cars. Using the cyclic stress-strain hysteresis loop equation and Neuber's Rule, the nominal strain histories obtained from tests were converted to local stress-strain responses at the fracture crack locations. Subsequently, the fatigue damage to the rear axles of both the target car and comparative car was calculated based on Morrow's mean stress correction model and Miner's linear cumulative damage rule. The frequency sweeping of vibration mode was performed on the car bodies and rear axles of both cars using electromagnetic exciters to establish the correlation between the vibration frequency of rear axle, the excitation frequency of PG durability road, and the body vibration frequency. The calculations of fatigue damage and the frequency sweep testing results indicated that the accumulated damage to the rear axle of the target car was primarily concentrated on the washboard road of the PG. On this road surface, the standard deviation of rear axle strain and the RMS value of acceleration were higher than those observed in the comparative car. At a test speed of 65 km/h, the excitation frequency of forced vibration caused by the washboard road was 24.1 Hz, which was close to the natural frequency of the vibration modes of target car's rear axle, leading to resonance. This resulted in the rear axle experiencing significant amplitude alternating stress, thereby causing the crack initiation of vibration fatigue.

Abstract: Commercial automotive radar systems for ADAS have relied on FMCW waveforms for years due to their low-cost hardware, simple signal processing, and established academic and industrial expertise. However, FMCW systems face challenges, including limited unambiguous velocity, restricted multiplexing of transmit signals, and susceptibility to interference. This work introduces a unified automotive radar signal model and reviews alternative modulation schemes such as PC-FMCW, PMCW, OFDM, OCDM, and OTFS. These schemes are assessed against key technological and economic criteria and compared with FMCW, highlighting their respective strengths and limitations.

Abstract: To reduce pollutants generated by automotive painting processes and improve coating efficiency, this study proposes a clean production method for automotive body painting based on an improved whale optimization algorithm from the perspective of "low-carbon consumption and emission-reduced production." A multi-level, multi-objective decision-making model is developed by integrating three dimensions of clean production: material flow (optimizing material costs), energy flow (minimizing painting energy consumption), and environmental emission flow (reducing carbon emissions and processing time). The whale optimization algorithm is enhanced through three key modifications: the incorporation of nonlinear convergence factors, elite opposition-based learning, and dynamic parameter self-adaptation, which are then applied to optimize the automotive painting model. Experimental validation using the painting processes of TJ Corporation's New Energy Vehicles (NEVs) demonstrates the superiority of the proposed algorithm over MHWOA, WOA-RBF, and WOA-VMD. Results show that the method achieves a 42.1% increase in coating production efficiency, over 98% exhaust gas purification rate, 18.2% average energy-saving improvement, and 17.9% reduction in manufacturing costs. This green transformation of low-carbon emission-reduction infrastructure in painting processes delivers significant economic and social benefits, positioning it as a sustainable solution for the automotive industry.

Abstract: Large amounts of high-quality data is required for the training of Artificial Intelligence (AI) models, which are indeed cumbersome to curate and perform quality assurance via human intervention. Moreover, models trained using erroneous data (human errors, data faults) can cause significant problems in real-world applications. This paper proposes an automated cleaning framework and quality assurance strategy for 2D object detection datasets. The proposed cleaning method was designed according to the ISO/IEC 25012 data quality standards, and uses multiple AI models to filter anomalies and missing data. In addition, it balances out the statistical unevenness in the dataset, such as the class distribution and object size distribution. Thereby ensuring the quality of the training dataset and examining the relationship between the amount of data required for enhanced performance in terms of detection. The experiments were conducted using popular datasets for autonomous driving, including KITTI, Waymo, nuScenes and publicly available datasets from South Korea. An automated data cleaning framework was employed to remove anomalous and redundant data, resulting in a reliable dataset for training. The automated data pruning and assurance system demonstrated the ability to substantially decrease the time and resources needed for manual data inspection.

Abstract: This publication aims to present the preliminary results of research on an innovative device designed to support the rehabilitation of drivers with neurological disorders, developed as part of a multidisciplinary project. The device was designed for individuals recovering from neurological diseases, injuries, and COVID-19-related complications, who experience difficulties with coordination and the speed of performing motor exercises. Its goal is to improve the quality of life for patients and increase their chances of safely driving vehicles, which also contributes to the safety of all road users. The device allows for controlled upper limb exercises using a diagnostic module, exercise program, and biofeedback system. The main component is a mechatronic driving simulator, enhanced with dedicated software to support the rehabilitation of individuals with neurological disorders and older adults. Through driving simulations and rehabilitation tasks, patients perform exercises that improve their health, facilitating a faster recovery. The innovation of the solution is confirmed by a submitted patent application, and preliminary research results indicate its effectiveness in rehabilitation and improving mobility for individuals with neurological disorders.

Abstract: Accurate water body detection is essential for autonomous navigation and 1 operational planning of Unmanned Surface Vehicles (USVs). To address model adaptability 2 to ambiguous boundaries caused by diverse scenarios and climatic conditions, this study 3 proposes GFANet (Global-Local Feature Attention Network) for real-time water surface 4 semantic segmentation of camera-captured images. First, a Global-Local Feature (GLF) 5 Extraction module is proposed, integrating a self-attention-based local feature extractor 6 and a multi-scale global feature extractor for parallel feature learning, thereby enhancing 7 hierarchical feature representation. Second, a Gated Attention (GA) Module is designed 8 with a dual-branch gating mechanism to implement noise suppression and efficient low- 9 level feature utilization. The method was validated on three publicly available datasets in 10 relevant domains.Experimental results on the Riwa dataset show GFANet achieves state-of- 11 the-art segmentation performance (4.41M parameters, 7.15 GFLOPs) with mIoU 82.29% 12 and mPA 89.49%. Comparable performance metrics were obtained on the USVInland and 13 WaterSeg datasets.Additionally, GFANet achieves 154.98 FPS processing speed, meeting 14 real-time segmentation requirements. Experimental results verify that GFANet achieves an 15 optimal balance between high segmentation accuracy and real-time processing efficiency.

Abstract: This study investigates the application of the Taguchi method for optimizing corrosion management strategies in biodegradable materials used in electric vehicles (EVs). As the automotive industry shifts toward sustainable practices, the use of biodegradable materials presents a promising alternative to traditional plastics and metals. However, these materials face significant challenges related to corrosion, which can compromise their performance and longevity in EV applications. The research employs a systematic approach, utilizing the Taguchi method to design experiments that evaluate the effects of various factors—including material type, manufacturing technique, and environmental conditions—on corrosion resistance. Biodegradable materials such as Polylactic Acid (PLA), Polyhydroxyalkanoates (PHA), and biocomposites were selected for analysis, and corrosion tests, including salt spray and electrochemical assessments, were performed to determine their durability. Results indicate that biocomposite materials exhibit superior corrosion resistance compared to pure biodegradable polymers, particularly when optimized through advanced manufacturing techniques. The findings also reveal that the Taguchi method effectively identifies optimal conditions that enhance the corrosion resistance of biodegradable materials. This research contributes to the growing body of knowledge on sustainable automotive materials and offers practical recommendations for manufacturers seeking to improve the durability and environmental impact of EV components. The insights gained from this study pave the way for future innovations in biodegradable materials and their integration into the automotive industry.

Abstract: The automotive 12 V power net is undergoing significant transitions driven by increasing power demand, higher availability requirements, and the aim to reduce wiring harness complexity. These changes are prompting a transformation of the power net architecture. To understand how future power net topologies will influence component requirements, electrical simulations are essential. They help to analyze the transient behavior of the future power net, such as under- and over-voltages, over-currents, and other harmful electrical phenomena. Accurate parametrization of the simulation models is crucial to obtain reliable results. This study focuses on the wiring harness, specifically its resistance and inductance, as well as the loads within the low-voltage power net, including their power profiles and input capacities. The parameters for this study were derived from vehicle measurements in three selected battery electric vehicles from different segments and enriched by virtual vehicle analyses. As a result, an extensive database of vehicle parameters was created, which is presented in the paper and can be used for power net simulations. In a next step, the collected data can be utilized to predict the parameters of various configurations in a zonal architecture setup.

Abstract: Vehicle accidents, particularly passenger vehicle-to-passenger vehicle (PV-PV) collisions, cause significant property damage and driver injuries, resulting in substantial economic losses and health risks. Most existing studies focus on macro-level predictions, such as accident frequency, but lack detailed collision-level analysis, which limits precise severity prediction. This study investigates various accident-related factors, including environmental conditions, vehicle attributes, driver characteristics, pre-crash scenarios, and collision dynamics.Using data from the National Highway Traffic Safety Administration’s (NHTSA) Crash Report Sampling System (CRSS) and Fatality Analysis Reporting System (FARS), the dataset was balanced by integrating complementary severity-level data with random over-sampling and under-sampling techniques. The Minimum Redundancy Maximum Relevance (mRMR) algorithm was used for feature selection to minimize redundancy and identify key features.Five advanced machine learning models were employed for severity prediction, with Extreme Gradient Boosting (XGBoost) achieving the best performance: 84.9% accuracy, 84.85% precision, 84.90% recall, and an F1-score of 84.87%. SHapley Additive exPlanations (SHAP) analysis was employed to interpret the model and conduct a deep analysis of accident features, including feature importance, dependencies, and their combined effects on severity prediction.The findings highlight the effectiveness of interpretable machine learning in understanding accident severity and identifying key factors influencing driver injuries.

Abstract: This study introduces a novel state estimation framework that incorporates Deep Neural Networks (DNNs) into Moving Horizon Estimation (MHE), shifting from traditional physics-based models to rapidly developed data-driven techniques. A DNN model with Long Short-Term Memory (LSTM) nodes is trained on synthetic data generated by a high-fidelity thermal model of a Permanent Magnet Synchronous Machine (PMSM), which undergoes thermal derating as part of the torque control strategy in a battery electric vehicle. The MHE is constructed by integrating the trained DNN with a simplified driving dynamics model in a discrete-time formulation, incorporating the LSTM hidden and cell states in the state vector to retain system dynamics. The resulting optimal control problem (OCP) is formulated as a nonlinear program (NLP) and implemented using the \texttt{acados} framework. Model-in-the-loop (MiL) simulations demonstrate accurate temperature estimation, even under noisy sensor conditions or failures. Achieving threefold real-time capability on embedded hardware confirms the feasibility of the approach for practical deployment. The primary focus of this study is to assess the feasibility of the MHE framework using a DNN-based plant model instead of focusing on quantitative comparisons of vehicle performance. Overall, this research highlights the potential of DNN-based MHE for real-time, safety-critical applications by combining the strengths of model-based and data-driven methods.

Abstract: A virtual model enables the study of reality in a virtual environment using a theoretical model, which is a digital image of a real model. The complexity of the virtual model must correspond to the reality of the evaluated system, becoming as complex as necessary nevertheless as simple as possible, allowing for computer simulation results to be vali-dated by experimental measurements. The virtual model of the autonomous vehicle was created using the CarMaker software package, which was developed by the IPG Auto-motive company and is extensively used in both the international academic community and the automotive industry. The virtual model simulates the real-time operation of a vehicle's elementary systems at the system level and provides an open platform for the development of virtual test scenarios in the Autonomous Vehicles, ADAS, Powertrain, and Vehicle Dynamics application areas. This model included the following virtual sen-sors: slip angle sensor, inertial sensor, object sensor, free space sensor, traffic sign sensor, line sensor, road sensor, object by line sensor, camera sensor, global navigation sensor, radar sensor, lidar sensor and ultrasonic sensor. Virtual sensors can be classified based on how they generate responses: sensors that operate on parameters derived from measurement characteristics, sensors that operate on developed modeling methods, and sensors that operate on applications.

Abstract: In recent years, significant progress has been made in multi-source navigation data fusion methods, driven by rapid advancements in multi-sensor technology, artificial intelligence (AI) algorithms, and computational capabilities. On one hand, fusion methods based on filtering theory, such as Kalman Filtering (KF), Particle Filtering (PF), and Federated Filtering (FF), have been continuously optimized, enabling effective handling of non-linear and non-Gaussian noise issues. On the other hand, the introduction of AI technologies like deep learning and reinforcement learning has provided new solutions for multi-source data fusion, particularly enhancing adaptive capabilities in complex and dynamic environments. Additionally, methods based on Factor Graph Optimization (FGO) have also demonstrated advantages in multi-source data fusion, offering better handling of global consistency problems. In the future, with the widespread adoption of technologies such as 5G, the Internet of Things, and edge computing, multi-source navigation data fusion is expected to evolve towards real-time processing, intelligence, and distributed systems. So far, Fusion methods mainly include optimal estimation methods, filtering methods, uncertain reasoning methods, Multiple Model Estimation (MME), AI, and so on. To analyze the performance of these methods and provide a reliable theoretical reference and basis for the design and development of a multi-source data fusion system. This paper summarizes the characteristics of these fusion methods and the corresponding adaptation scenarios. These results can provide references for theoretical research, system development, and application in the fields of autonomous driving, unmanned vehicle navigation, and intelligent navigation.

Abstract: This paper proposes a method for predicting the wear and tear of automotive consumables and wear parts based on their quality, driving style and road conditions, integrating IoT sensors for real-time monitoring. The developed model aims to analyze historical and real-time data to estimate the components' lifetime and provide personalized maintenance recommendations. Using noise, vibration and temperature sensors, Machine Learning algorithms and statistical analysis, the model can optimize the predictive maintenance strategy, thus reducing unforeseen costs and extending the vehicle's lifespan. This method overcomes the limitations of standardized manufacturer warranties and offers a practical and adaptable solution for the automotive industry.

Abstract: Turbocharging of hydrogen fuel cell electric vehicles (FCEVs) has recently become a prominent research area, aiming to improve FCEV efficiency and viability to help decarbonise the transport sector. This work compares the performance of electrically assisted variable-geometry-turbocharging (VGT) with a fixed-geometry turbocharger (FGT) by optimising both the sizing of the components and their operating points, ensuring both designs are compared at their respective peak performance. A MATLAB-Simulink reduced-order model is used first to identify the most efficient components which match to the fuel cell air-path requirements. Maps representing the compressor and turbines are then evaluated in a 1D flow model to optimise cathode pressure and stoichiometry operating targets for net system efficiency, using an accelerated genetic algorithm (A-GA). Good agreement was observed between the two models’ trends, with negligible difference in system efficiency and modelled shaft speed under optimised conditions. However, a sensitivity study demonstrates significantly higher efficiency when operating at non-ideal flows and pressures for the VGT when compared to the FGT, suggesting that VGTs may provide higher level of tolerance under variable environmental conditions such as ambient temperature, humidity, and transient loading. Overall, it is concluded that the efficiency benefits of VGT are marginal, and therefore not necessarily significant enough to justify the additional cost and complexity that they introduce.

Abstract: In recent years, synthetic aperture radar (SAR) technology has been increasingly explored for automotive applications. However, automotive SAR images generated via the matched filter (MF) often exhibit challenges such as noisy backgrounds, sidelobe artifacts, and limited resolution. Sparse regularization methods ave the potential to enhance image quality. Nevertheless, conventional unweighted ℓ1 regularization methods struggle to address cases with radar cross section (RCS) distributed over a wide dynamic range, often resulting in insufficient sidelobe suppression, amplitude distortion, and inconsistent super-resolution performance. In this paper, we propose a novel reweighted regularization method, termed Multi-Segment-Reweighted Regularization (MSR), for automotive SAR image restoration. By introducing a novel weighting scheme, MSR localizes the global scattering point enhancement problem to the mainlobe scale, effectively mitigating sidelobe interference. This localization ensures consistent enhancement capability independent of RCS variations. Furthermore, MSR employs multi-segment regularization to constrain amplitude within the mainlobes, preserving the characteristics of the original response. Correspondingly, a new thresholding function, named Thinner Response Undistorted THresholding (TRUTH), is introduced. An iterative algorithm for enhancing automotive SAR images using MSR is also presented. Real data experiments validate the feasibility and effectiveness of the proposed method.

Abstract: This study aims to describe the advantages of electrifying regular-route buses. The results of a survey on the speed profile of buses operating on actual routes are presented here. Firstly, we focus on the acceleration / deceleration at the starting / stopping stops specifically for regular-route buses and obtain the following information: I. Starting acceleration from a bus stop is particularly strong in the second half of the acceleration process, being suitable for motor-driven vehicles. II. The features of the stopping deceleration at a bus stop are “high intensity” and “low dispersion,” with the latter enabling the refinement of regenerative settings and significantly lowering electricity economy during electrification. And we compare the speed profile of a fuel cell bus with those of a diesel bus and obtain the following information: III. Motor-driven vehicles have “high acceleration performance” and “no gear shifting” which facilitate the high-intensity stop–start acceleration operation unique to regular-route buses. By calculating and analyzing the jerk amount, we could quantitatively demonstrate the comfortable driving experience while riding on this type of bus where there is no shock due to gear shifting. IV. While the “high acceleration performance” of motor-driven vehicles produces “individual differences in the speed change patterns,” this does not translate to “individual differences in electricity consumption” owing to the characteristics of this type of vehicle. With engine-driven vehicles, measures, such as “slow acceleration” and ”shift up early” are strongly encouraged to realize eco-driving, and any driving style that deviates from these measures is avoided. However, with motor-driven vehicles, the driver does not need to be too concerned about the speed change patterns during acceleration. This characteristic also suggests a benefit in terms of the electrification of buses.

Abstract: This study presents a methodology to develop an energy consumption model for electric vehicles based on dynamic vehicle and environmental data. Particular attention is given to analyzing the impact of ambient temperature on energy consumption modeling. To improve model accuracy, temperature and vehicle acceleration data were clustered using the K-Means method. As a result, four energy consumption models were created, each corresponding to a specific data cluster. This approach achieved strong validation results, with an R² value of 0.84 and a MAE ranging from 0.75 to 1.23 Wh, which is a satisfactory outcome given the large volume of input data. These models are designed for microscale energy consumption prediction, using vehicle speed and acceleration as input parameters.

Abstract:

This study introduces an enhanced Forest PSO-GA algorithm for forest fire monitoring, integrating Particle Swarm Optimization (PSO), Genetic Algorithm (GA), and forest wind dynamics, while considering the impact of terrain variations on energy con-sumption, into an adaptive search framework. By incorporating wind-driven fire propagation and smoke diffusion models into a cellular automata simulation platform, the algorithm effectively evaluates its performance and further accounts for elevation and energy consumption. This enables more accurate simulation of fire and smoke spread, ensuring efficiency and sustainability in remote forest areas. Simulations using data from the Harbin Liangshui Forest show that the enhanced Forest PSO-GA out-performs APSO, AFSA, and PSO-PID in search speed by 91.34%, 340.89%, and 52.21%, respectively. It achieves an average localization accuracy of 9 meters (±1.2 meters), which is sufficient for the precise deployment of fire-extinguishing devices. The algo-rithm also reduces the search area by 35.4-72.3% and converges within 50 iterations 80% of the time, representing a 28.7% efficiency gain over PSO-PID. Additionally, the algo-rithm boasts a success rate of 94.3% and a 61.8% improvement in wind resistance, ef-fectively supporting pre-disaster warnings and early fire detection. These advance-ments significantly enhance fire detection accuracy, reduce the burden on forest fire prevention efforts, and improve precision firefighting and ecological recovery capabil-ities, offering a highly efficient and reliable solution for forest fire management.

Abstract: This research paper investigated the critical role of real-time automated awareness and the handling of exceptions in developing autonomous driving technologies. Focusing on the automated driving systems deployed by Tesla, Waymo, and NVIDIA, we explored the integration of real-time data processing, sensor technology, and AI algorithms essential for navigating complex driving scenarios. Exception-handling mechanisms were studied, highlighting their significance in enhancing vehicle safety and reliability. Through adaptive learning and software updates, these autonomous systems continuously improved their capabilities to address unexpected challenges, increasing their decision-making efficacy and reducing risks. The study also touched upon the implications of AI in autonomous driving, addressing ethical considerations and the importance of human-computer interaction for user trust and safety. By synthesizing recent research findings, this paper presented an overview of the present advancements and potential future directions in autonomous driving, emphasizing the need for ongoing innovation in real-time awareness and exception management.

Abstract:

Due to the growth of pollutant emissions caused by the burning of fossil fuels (CO2) produced by internal combustion vehicles, the automotive industry aims to increasingly reduce fuel consumption in vehicles. In the present investigation, we analyzed how the different driving modes of an HEV influence the rate of fuel consumption per kilometer traveled during urban trips in the city of Guayaquil, Ecuador. A 2022 Toyota Corolla Hybrid vehicle was used for the study and the SAE J1321:2012 Standard was used, which provides a standardized method for fuel consumption evaluation. Using the OBDLINKMX+ device for real-time data collection during the trip with the help of the GPS Speedometer application. We took into consideration the routes and hours where there is the highest vehicle demand within the city of Guayaquil (North, Center and West). The driving modes analyzed were (Normal, Eco, Power), the results were processed in the Minitab statistical software, showing that the fuel consumption rates during the established tests were: 0.013536 Gal/km (Normal), 0.012472 Gal/km (Eco), 0.014658 Gal/km (Power), providing information on how the factors: travel time, average speed, travel distance, as well as traffic conditions and the usual behavior of a driver, affect fuel consumption.