Thermal runaway in lithium-ion batteries presents a critical safety challenge for electric vehicles. While time-temperature profiles offer valuable insights into thermal runaway behavior, most existing studies narrowly focus on characteristic temperatures and heating rates, often overlooking essential dynamic details. Moreover, current evaluation methods lack objectivity and fail to integrate multiple contributing factors. This study proposes a phased analysis approach that extracts time-temperature indicators and applies an entropy-weighted TOPSIS (technique for order preference by similarity to ideal solution) model to objectively assign weights to different stages across 10 experimental cases. Results indicate that the induction period contributes the most to safety risks (34.3 %), followed by the post reaction period (28.7 %). Tests on cylindrical NCM cells at 50 % state of charge (SOC) under thermal abuse conditions reveal that thermal runaway triggered at lower SOCs may result in incomplete internal reactions, thereby prolonging high-temperature cooling phases-an important finding for mitigating thermal propagation. The proposed objective ranking system enables more comprehensive assessments of thermal runaway and highlights the importance of phase-specific analysis in battery safety management.

With the escalation of environmental issues, electric vehicles (EVs) powered primarily by lithium-ion batteries (LIBs) have become a significant alternative to traditional fuel vehicles. However, the pursuit of high energy density and high-rate performance has led to increasingly severe thermal safety issues related to thermal runaway (TR), which significantly hinders the further promotion of electric vehicles. To address this, this paper first reviews the mechanisms associated with thermal runaway, including its triggers, development, and propagation, which aids in a deeper understanding of thermal runaway. Subsequently, based on these mechanisms, the current advanced thermal runaway protection methods are summarized from both component-level and system-level perspectives. At the component level, enhancing the thermal safety performance of batteries through material reinforcement is the most direct approach, as these materials are the main participants in the side reactions of thermal runaway. However, when abuse scenarios occur, lithium-ion batteries still face the risk of thermal runaway; therefore, system-level protective methods are also necessary. On one hand, advanced thermal management technologies can be utilized to prevent the accumulation of heat within the battery before thermal runaway occurs, thereby reducing the likelihood of triggering thermal runaway. On the other hand, after a thermal runaway event occurs, suppression and extinguishing measures can be employed to inhibit its development and propagation, thus mitigating the hazards posed by thermal runaway. This paper constructs a comprehensive thermal safety protection system through a thorough review of the latest research findings in the field of thermal safety protection, providing maximum safety assurance for batteries operating at high rates.

Benefiting from its over-actuated characteristics, the distributed vehicle offers enhanced maneuverability and controllability under extreme conditions compared to conventional architectures. Integrated motion control serves as a crucial means to resolve conflicts between actuator functions and achieve dynamic harmony, thus enhancing the accuracy and stability of the trajectory tracking of autonomous vehicles. High-precision motion control relies on the accurate dynamics model. However, the mechanism-based control strategy suffers from structural inaccuracies, environmental disturbances, and parameter perturbations, leading to performance degradation and poor robustness. In this paper, a fusion control strategy is proposed that combines the generalization of the mechanism model with the precision of data-driven learning to predict unmodeled dynamics. Specifically, within the framework of offset free model predictive control (OF-MPC), Bayesian optimization (BO) is introduced to identify the key parameters of the mechanism nominal model under typical coupling motion conditions, and a states-coefficients joint mismatch estimator (SCJME) is trained by physics-informed neural network (PINN) with mechanism constraints introduced in the training process to ensure the stability and convergence of the estimator, which enables the SJCME to predict the unmodeled dynamics and compensate for nominal model mismatch during rolling optimization. In particular, the dual extended Kalman filter (DEKF) is adopted in the construction of SCJME, which adjusts the fusion coefficient online based on the states-observation residuals to mitigate performance degradation in platform migration. In comprehensive scenarios, hardware-in-loop (HIL) tests have validated the effectiveness of the proposed strategy, demonstrating improved precision and stability in coupled motion tracking, along with enhanced robustness during heterogeneous platform migrating.

For humanu2012machine co-driving vehicles, the dynamic complexity of urban scenarios and the uncertainty of driving behavior impose stringent demands on the accuracy and generalization capability of ego vehicle trajectory prediction. Current research predominantly relies on multivehicle interaction information or single-scenario settings, overlooking the inherent dynamic correlation between driver and vehicle. This results in prediction models struggle to adapt to complex urban environments. To address this, this study proposes a trajectory prediction framework based on driver-vehicle coupling feature encoding, enabling precise capture of vehicle trajectory evolution patterns under driver manipulation across diverse urban scenarios. The framework employs bidirectional long short-term memories (LSTMs) to perform temporal encoding on historical driver-vehicle coupling features and future road geometric features. Combined with an attention mechanism, it generates context vectors integrating temporal features and manipulation details, ultimately using LSTMs to recursively produce multistep prediction results. Validation using diverse urban scenarios data from a dynamic driving simulatordemonstrates that our prediction framework achieves precise trajectory prediction in typical scenarios such as lane changing, turning, and roundabout navigation, while exhibiting robust stability and generalization capabilities.

With the rapid development of autonomous driving technology, unmanned ground vehicles (UGVs) are gradually replacing humans to perform tasks such as reconnaissance, target tracking, and search in special scenarios. Omnidirectional mobility based on rapid adjustment of vehicle heading posture enhances the applicability of UGVs in specialized scenarios. Omnidirectional mobility signifies the capability for rapid adjustments to the vehicle’s heading angle, longitudinal velocity, and lateral velocity. Traditional vehicles are constrained by the limitations of under-actuation, which prevents active regulation of lateral movement. Instead, they rely on the coordinated regulation of longitudinal and yaw movements, failing to meet the requirements for omnidirectional mobility. Distributed vehicles featuring steering distributed between the front/rear axles and four-wheel independent drive leverage the over-actuation advantages provided by multi-actuator coordinated control, making them particularly suitable for omnidirectional mobility at large sideslip angles. This feature enables the UGVs to achieve rapid adjustment of vehicle heading posture. However, existing control strategies centered on stabilizing yaw rate and suppressing sideslip angles cannot adapt to the decoupling control requirements of such platforms. Additionally, the strong coupling characteristics between actuator subsystems further exacerbate control difficulties. To this end, this article proposes a full-state decoupling motion control strategy, the nonlinear model is locally linearized at each equilibrium point of the vehicle, and a set of equilibrium state models is derived. The validity of this local linearization method is verified through phase diagram analysis and modal analysis. The Bayesian optimization (BO) algorithm is then employed to optimize and identify the cornering stiffness of the front/rear axles at each equilibrium point in these locally linearized models, thereby enhancing the characterization ability of the linear model for the nonlinear dynamic model at the corresponding equilibrium points. Subsequently, a full-state decoupling motion controller is designed by integrating the model predictive control (MPC) algorithm. Finally, the controller presented in this article is employed on the distributed vehicle experiment platform (DVEP). The experimental results demonstrate that in two drift-like scenarios with different sideslip angles, compared with the baseline controller, the path tracking error of this method is reduced by more than 13%, and the sideslip angle tracking error is reduced by more than 12%.

In order to reduce the low-frequency, high-intensity vibration of a commercial vehicle cab and improve ride comfort, this paper proposes a multi-objective optimization method for cab parameters based on an improved particle swarm optimization (PSO) algorithm. The novelty of this work lies in the integration of a Kent chaotic map to initialize the particle swarm, which enhances population diversity and uniformity, thereby improving the algorithm’s ability to avoid local optima and increasing convergence stability. A multi-body dynamics model, constructed in ADAMS, is utilized to determine the cab’s key suspension parameters (stiffness, damping, and position). A multi-objective optimization model is then established with the objectives of minimizing the root mean square (RMS) value of vertical cab acceleration, minimizing the comprehensive weighted acceleration, and maximizing the decoupling degree of rigid body vibration modes. The improved PSO algorithm is used to solve the model and obtain the optimal cab parameter set. Experimental results, validated on a physical JH6 commercial vehicle under three distinct road conditions, show that this method effectively optimizes the cab parameters, leading to a measured ride comfort improvement of up to 82.5

Ensuring the safety of Autonomous Vehicles (AVs) continues to present a significant challenge. In order to achieve broader acceptance, it is essential to ensure that autonomous driving systems consistently make safe decisions in all traffic scenarios and are validated for safety compliance. This paper proposes a Predicted Occupancy Map-Based Online Safety Verification and Resilient Motion Planning (POM-SVP) method as a safety layer for current motion planning algorithms. The method aims to verify whether the planned trajectories of autonomous driving systems meet safety requirements and to provide resilient safe trajectories in critical scenarios. By combining explicit and implicit traffic rules, as well as kinematic models of vehicles and pedestrians, the proposed method effectively predicts the future over-approximated occupancy region of surrounding traffic participants to assess whether the intended trajectory of the autonomous vehicle presents any safety conflicts. Furthermore, our method utilizes a Transformer module to provide resilient, conflict-free safety trajectories with respect to the future occupancy of surrounding traffic participants, serving as backup switching solutions to ensure the vehicle always operates within a safe region. Experimental results indicate that the proposed method can significantly reduce the number of traffic accidents while maintaining vehicle motion continuity and intention consistency. By integrating our safety verification algorithm, we significantly outperformed previous research without integration on the widely recognized CARLA leaderboard, with the driving score improving by up to 119.13%, surpassing the current top-ranked entry.

Learning and simulating the decision processes of real-world human drivers is a key research direction in autonomous driving (AD). As the core of AD, existing decision systems typically face challenges in cross-scene generalization and decision interpretability: it requires understanding diverse dynamic driving scenarios and formulate transparent strategies that earn broad user trust. We propose VLM-Driver, a Vision-Language Model (VLM) framework with human-like chained driving decision thought, designed to progressively achieve global scene observation, high-level behavior planning, and low-level motion planning based on full-view driving videos, multi-round question queries and optional environmental perception information. Specifically, the full-view videos are resized to unified base resolution using the AnyRes strategy, and video features are extracted by a vision encoder. These video features are then mapped to the text embedding space via a vision-language projector and jointly fed into large language model (LLM) backbone with tokenized text embeddings. During this process, we introduce a Bilinear Interpolation method to efficiently compress the number of video features, ensuring an optimal balance between model performance and computational cost. Meanwhile, a dedicated motion head is designed to output refined future waypoints, improving the model's motion planning efficiency. Additionally, we construct a novel multimodal driving instruction dataset to support VLM-Driver training and introduce a decision-oriented training strategy to further enhance its chained reasoning capability. Extensive experiments show that VLM-Driver excels in scene observation and behavior planning, demonstrating high human-like consistency, and significantly outperforms existing baseline methods in motion planning, achieving SOTA performance. VLM-Driver also enable to handle unseen complex driving scenarios, exhibiting robust cross-scene generalization.

With the rapid proliferation of electric vehicles, particle emissions from lithium-ion batteries during thermal runaway (TR) pose growing health and environmental concerns. While previous studies focused on thermal abuse, the characteristics of particle emissions under mechanical abuse (e.g., nail penetration) remain underexplored, particularly regarding state of charge (SOC) effects. Here, we investigate SOC-dependent particle emissions from 18650 LIBs (lithium nickel cobalt aluminum oxide cathode) during nail penetration. Utilizing a standardized experimental platform for triggering TR combined with a particle emissions collection system, we employed laser particle size analysis, scanning electron microscopy-energy dispersive spectrometry and inductively coupled plasma mass spectrometry to quantitatively analyze the size distributions, morphological characteristics, and major metal compositions of particle emissions. The results reveal that increasing SOC leads to a multiplicative growth in the median diameter of particle emissions, while simultaneously promoting more irregular and porous morphological features and doubling the concentrations of major metals (Nickel/ Cobalt/ Copper). The particle emissions primarily originate from the thermal decomposition of battery components coupled with subsequent redox reactions. This study provides a theoretical basis for battery safety design and emergency protection against TR.

Motion planning for autonomous driving requires consideration of close interactions between the ego-vehicle and surrounding traffic participants, resulting in a synergistic relationship between path planning and motion prediction tasks. However, existing approaches often rely on static adjacency graphs or decoupled pipelines, making it difficult to capture dynamic interagent dependencies and scene variations. This leads to limited long-horizon predictive capability and planned trajectories that lack collaborative responsiveness and temporal consistency. Especially in dense urban traffic, many multiscale or interaction-aware methods remain prediction-centric and lack a unified mechanism to jointly model multiscale spatiotemporal information, causal constraints, and planning objectives, making it challenging to balance long-term behavioral consistency with short-term risk adaptation. To address this, we propose ACP-MSP, a multiscale unified prediction-planning framework that enhances scene understanding and decision-making through deep integration of prediction and planning. First, we construct a dynamic topology interaction network to unify the representation of multiagent trajectories and environmental elements, and generate multimodal probabilistic trajectories through a multiscale hierarchical structure to improve long-term prediction capability. Second, we design a Transformer-based unified prediction-planning architecture that employs adaptive feature aggregation and cross-task attention to reduce information distortion and enhance collaborative decision-making. Finally, we introduce a causal self-attention mechanism that incorporates temporal masking constraints and scene-dynamic scoring based on obstacle distance and velocity changes, improving the temporal consistency of planned trajectories. Evaluations on the large-scale nuPlan and Val14 benchmarks show that ACP-MSP achieves state-of-the-art performance and demonstrates robust transferability across diverse driving styles.

Sideslip angle and velocity are key parameters for assessing vehicle stability during typical cornering and drifting maneuvers, but accurate measurement under varying roads is prohibitively expensive. Therefore, a diffusion-augmented Koopman Kalman filter (Diff-AKKF) is proposed in this article, enabling high-precision estimation of sideslip angle and velocity within and beyond the stability limits across different road surfaces. First, a conditional diffusion model with physical constraint is established to directly predict parameters rather than noise, enabling inference of equivalent dynamics parameters across various maneuvers to provide prior information on dynamics characteristics and state evolution to augment estimation. Subsequently, the Koopman Kalman filter is developed based on Koopman operator theory combined with deep learning, embedding parameters inferred from the diffusion model to estimate sideslip angle and velocity. Finally, real-vehicle experiments involving typical cornering and drifting maneuvers are conducted on both wet and dry asphalt roads, thoroughly validating the effectiveness, estimation accuracy, and robustness of the Diff-AKKF algorithm within and beyond the stability limits scenarios.

Semi-active suspension systems enhance ride comfort and handling performance by adaptively modulating damping characteristics. However, conventional model-based controllers often fail to maintain optimal performance under uncertain and time-varying vehicle conditions. This article proposes Bayesian Optimization–Tuned Proximal Policy Optimization with Non-Parametric Rewards (BO-NRPPO), a novel reinforcement learning (RL) framework that integrates Bayesian Optimization (BO) with Proximal Policy Optimization (PPO) and a non-parametric reward function (NRF). The proposed approach enables adaptive self-tuning, data-driven reward shaping, and uncertainty-aware policy learning. Moreover, a Trapezoidal Simple Moving Average (TSMA)–based reward normalization scheme is introduced to accelerate convergence and stabilize training. Simulation results across diverse driving scenarios demonstrate that BO-NRPPO outperforms the passive suspension, the classical Linear Quadratic Regulator (LQR), and PPO with parametric rewards. Specifically, compared to the passive suspension and the LQR baseline, BO-NRPPO achieves up to 6.63% and 5.14% improvements in handling stability, respectively. Concurrently, it delivers maximum enhancements of 46.96% and 42.55% in ride comfort over these two baselines. For real-world vehicle applications, this adaptive self-tuning capability significantly reduces the time-consuming manual calibration efforts typically required in chassis development. Furthermore, Hardware-in-the-loop (HiL) validation confirms its real-time applicability and robustness under uncertain driving conditions, highlighting its immense potential as a scalable intelligent suspension control solution.

Road rage is a major precipitant of traffic accidents in modern road transport, yet existing in-vehicle detection methods rely predominantly on behavioral or contextual cues, limiting their ability to capture underlying affective and cognitive mechanisms; meanwhile, electroencephalography (EEG) and peripheral physiological signals can provide "privileged" evidence, but their sensing burden hinders scalable deployment. We propose a Brain-Body-Vehicle Knowledge Distillation Network (BBV-KDNet), a teacher-student framework that integrates EEG, peripheral physiological signals, and Controller Area Network (CAN) bus vehicle-dynamics signals during training while enabling vehicle-dynamics-only inference. The multimodal teacher (BBV-KDNet-T) encodes each modality and unifies them via dynamic gated fusion and Transformer-based cross-modal interaction for fourlevel road-rage intensity estimation, and further outputs sample-level modality-contribution scores to support transparent interpretation. A lightweight student (BBV-KDNet-S) is trained with joint distillation to transfer both the teacher's decision boundaries and multimodal knowledge into a CAN-only model. We evaluate BBV-KDNet on a driving study with 24 participants under two conflict scenarios, constructing a synchronized EEG-physiology-CAN dataset with four-level road-rage labels. BBV-KDNet-T achieves balanced accuracies of 91.9% and 92.9% in Experiments A and B, respectively, while the CAN-only student improves mean accuracy over a CAN-only from-scratch baseline from 65.57% to 75.48% and from 69.97% to 79.93%, respectively. Calibration analysis supports using the student's outputs as continuous risk scores, while source localization links EEG patterns to plausible neural mechanisms. These results highlight artificial-intelligence-based multimodal learning and knowledge distillation for deployable road-rage detection in intelligent vehicles.

Nonlinear model predictive control (NMPC) algorithms have been widely used in autonomous vehicle trajectory tracking, yet their performance is primarily limited by the accuracy of the prediction model. This paper introduces a physics-informed neural network (PINN) dynamics model architecture that integrates a physics-based dynamics model into the neural network training process. The PINN optimizes all physical dynamics parameters, enforces the physical constraint to ensure parameter validity, and enhances model interpretability and fidelity. Furthermore, the PINN dynamics model is more practical by using accurately measurable wheel speed as inputs instead of torque through the coupled slip tire model. On this basis, a novel NMPC algorithm for trajectory tracking is proposed, which optimally balances the prediction horizon and distance by integrating the short and long horizons to ensure the tracking performance while considering the real-time computational efficiency. Finally, the real-vehicle experiment is conducted on a distributed drive electric vehicle (DDEV) to verify the algorithm's effectiveness with good real-time performance. The PINN dynamics model exhibits superior accuracy to the physics-based dynamics model, significantly enhancing the NMPC trajectory tracking capabilities.

Partially automated driving systems still rely on human drivers to monitor the environment and resume control when needed. Due to reduced vigilance and situation awareness during automation, drivers may misjudge the level of risk and the system’s capability to handle sudden conflict scenarios, which can undermine overall safety. Effectively communicating risk-predictive information generated by the automation through the Human–Machine Interface (HMI) can help drivers better understand system behavior and adjust their trust accordingly. In this study, we conducted a 3 × 2 × 2 within-subjects experiment involving intersection conflict scenarios (HMI: baseline vs. risk-alert vs. multimodal alert; risk level: low vs. high; turning direction: left vs. right). Thirty participants took part in the simulation-based experiment, with psychological measures, peripheral biosignals, neural activity, and driver-initiated takeover behavior recorded throughout the experiment. Results showed that risk-predictive information amplified the difference in drivers’ subjective risk evaluations between high-risk and low-risk conditions, and promoted more effective cognitive control in high-risk scenarios. The multimodal alert HMI (integrating visual and auditory risk prediction information) increased situational trust and reduced takeover frequency. High-risk and right-turn conditions led to higher perceived risk, lower trust, and more frequent takeovers. These findings provide new insights into how risk-predictive HMI designs influence drivers’ performance in partially automated driving, contributing to safer and more effective human–automation interaction.

End-to-end autonomous driving, which directly maps raw sensor inputs to low-level vehicle controls, is an crucial part of Embodied AI. Despite successes in applying Multimodal Large Language Models (MLLMs) for high-level traffic scene semantic understanding, it remains challenging to effectively translate these conceptual semantics understandings into low-level motion control commands and achieve cross-scene driving generalization and consensus. We propose Sce2DriveX, a human-like chain-of-thought (CoT) driving reasoning MLLM framework, designed to achieve progressive learning from multi-view scene understanding to behavior analysis, motion planning, and vehicle control driving process. Sce2DriveX utilizes multimodal joint learning of local scene videos and global Bird's Eye View (BEV) maps to deeply understand long-range spatiotemporal relationships and road topology, enhancing its 3D dynamic/static scene perception and reasoning capabilities and achieving cross-scene generalization. Meanwhile, it reconstructs the implicit cognitive chain inherent in human driving, further enhancing the consensus between autonomous driving and human thought. To improve model performance, we construct the first comprehensive Visual Question Answering (VQA) driving instruction dataset, which tailored for 3D spatial understanding and long-axis task reasoning, and introduce a task-oriented three-stage training pipeline to support supervised fine-tuning. Extensive experiments demonstrate that Sce2DriveX achieves state-of-the-art performance across tasks from scene understanding to end-to-end driving, as well as robust generalization in handling diverse driving scenes on the CARLA Bench2Drive benchmark.

Motor imagery (MI) based brain–computer interfaces (BCIs) enable active device control through electroencephalography (EEG), offering contactless human–machine interaction. However, realistic MI-EEG decoding is challenged by non-uniform electrode coupling, long-range rhythmic dynamics, unstable second-order statistics, and blurred boundaries among fine-grained motor intentions. These challenges require topology-aware, statistically stable, and boundary-discriminative EEG representations, which remain insufficiently addressed in realistic decoding scenarios. To address these challenges, we propose ASDL-EEG, a coordinated representation learning framework for robust MI-EEG decoding. First, ASDL-EEG introduces a topology-aware first-order encoder that performs asymmetric electrode-axis spatial aggregation with multi-scale receptive fields and dilated temporal modeling to capture non-uniform spatial dependencies and long-range MI rhythms. Second, the Log-Diagonal Riemannian (LDR) extractor constructs aligned log-diagonal covariance descriptors as a compact, stability-oriented alternative to high-dimensional full covariance representations. To further enhance boundary discrimination, we propose RPA-v2, a sample-specific hard-negative prototype alignment method for fused first- and second-order EEG embeddings enlarging the margin between the target prototype and the most confusing non-target prototype. We further construct CW-MI-5, to the best of our knowledge, the first five-class MI-EEG dataset for intelligent-cockpit car-window control, serving as a naturalistic benchmark with synchronized EEG sensing and cockpit interaction cues. Experiments on CW-MI-5 and BCI Competition IV-2a show superior accuracy and Cohen’s kappa, validating ASDL-EEG in both cockpit interaction and standard MI benchmark settings.

The driver’s situation awareness (SA) at the moment of a takeover request (TOR) is essential for safe control transition in conditionally automated driving. To address limitations in existing methods for SA quantification and prediction, this study proposes an interpretable framework based on the analysis of SA’s formation process and critical role during takeovers. First, an eXtreme Gradient Boosting (XGBoost) model was constructed to predict takeover performance (TOP). SHapley Additive exPlanations (SHAP) was employed to provide local interpretability and identify the contribution of readily measurable SA Level 1 to TOP, based on which SA Level 2 and 3 were operationalized, thereby enabling the continuous quantification of overall SA. Subsequently, another XGBoost model was developed to predict the quantified SA, and SHAP’s global interpretability was used to optimize model performance and analyze the key features’ main and interaction effects on SA. A takeover simulation experiment was conducted to construct the necessary dataset. Results indicated that the quantified SA was able to match the reasonable distribution patterns of SA across different takeover results. The XGBoost model outperformed four comparison models in SA prediction. In addition, the key features’ main and interaction effects on SA were broadly supported by existing studies, confirming the rationality and effectiveness of the SA prediction model. This research provides support for driver state monitoring and the formulation of takeover strategies in conditionally automated driving.

Driving anger is strongly associated with aggressive driving and elevated crash risk. However, continuous modeling of graded anger-related affective states from physiological signals under controlled on-road exposure remains underexplored. In this study, we investigated driving-anger-related affective states elicited in an on-road Wizard-of-Oz (WoZ) paradigm, in which 24 licensed participants sat in the front passenger seat, imagined themselves as the driver, and were exposed to scripted traffic-conflict maneuvers executed by a trained safety driver. Using synchronized 32-channel electroencephalography (EEG), self-reports, and contextual information, we constructed a four-level anger-state dataset. Building on this dataset, we introduce TRIDENT, a multi-task deep learning model designed to simultaneously recognize four levels of anger states, estimate continuous anger severity (0-100), and predict future high-anger states. TRIDENT integrates multi-scale temporal convolutions, brain-network representations, and sequence modeling to capture complementary spatiotemporal patterns of anger dynamics. Experimental results show that TRIDENT significantly outperforms representative baseline EEG emotion recognition models, achieving up to 85% accuracy in four-class anger-state classification and 87% accuracy in predicting future high-anger states. Scalp topographies and cortical source localization analyses further reveal anger-level-dependent changes in prefrontal, temporal, and limbic brain networks. These findings provide a physiologically grounded perspective on anger-related neural dynamics under controlled on-road conflict exposure and have implications for emotion-aware in-vehicle interfaces and personalized intervention design. Code will be made available upon acceptance at: https://github.com/tianyaz719/TRIDENT.

A game-based cooperative steering control (GCSC) approach is introduced to facilitate effective collaboration between human drivers and automation, incorporating the neuromuscular delay inherent in human responses. In this framework, a dynamic coordination between driver and automation goals is achieved through the establishment of a game equilibrium in instances of driver and automation conflict. In response to the challenges posed by frequent modifications in driving weights and their subsequent burden on human drivers, this article proposes a strategy that integrates fixed initial weights with dynamic adjustments to driver-automation driving weights. Moreover, a comprehensive evaluation method including subjective and objective evaluation indexes is proposed. Different drivers are invited to perform virtual driving experiments, and the experimental results are analyzed by the proposed evaluation method. It is concluded that the driver's driving weight should be kept at a high level during cooperative steering control when the driver's intention cannot be perfectly obtained, and the determination of the driving weight should also consider the driver's driving skills.