Distributed electrically-driven heavy-duty vehicles achieve compound steering through the differential between the steering assist motor and the wheelend drive motor. By means of coordinated control of multiple motors, various active safety control functions are realized and the driver's operational burden is reduced. For the driving safety issues brought about by the failure of the drive motor and the steering assist motor, in this paper a faulttolerant control strategy encompassing mode-switching and faulttolerant torque distribution is proposed. The proposed modeswitching strategy, based on the vehicle pose information, introduces in the yaw rate residual function as the switching condition for the faulttolerant mode. The proposed faulttolerant torque distribution strategy takes into account of the output redundancy and the vehicle's stability to solve for the target output torque of the drive motor and the steering assist motor. Finally, a hardware-in-the-loop simulation platform is established to verify the effectiveness and real-time performance of the control strategy.

Establishing an accurate air spring model is key and crucial for analyzing the vibration characteristics of air suspension electric buses. For the variation in the characteristics of air springs under different load, a comprehensive model of the air spring with dynamically adjustable parameters is proposed, considering the effect of rubber airbag force and changes in payload, taking a membranetype air spring as the research object. Key parameters of the rubber airbag are identified through mechanical experiments, and the accuracy and effectiveness of the proposed model are verified. Based on the comprehensive air spring model, a 14seat, 21degreeoffreedom electric bus dynamics model is established. The model's validity is verified through simulation comparison with a CarSim model with identical parameters. Subsequently, the influence of air spring nonlinear characteristics, vehicle speed, road roughness, and passenger distribution on the dynamic performance of the electric bus system is analyzed. The study shows that the proposed comprehensive air spring model can dynamically adjust its parameters in response to variation in load and road excitation. The hysteretic mechanical characteristics of the air spring cannot be ignored. Compared with linear models, thermodynamic models without considering hysteresis, and equivalent air spring model, the comprehensive air spring model significantly reduces suspension deflection, with reduction of 22.95%, 42.13%, and 18.20%, respectively. Increase of vehicle speed, lower road quality, and uneven passenger distribution negatively affect the ride comfort of the electric bus, with discomfort increasing for passengers seated farther from the bus's center of gravity.

There are many investigated parameters in the powertrain mounting system (PMS) of electric vehicles, and it involves multiperformance design. For the problem that the traditional singleoutput sensitivity analysis is difficult to accurately evaluate the influence of system parameters on the system comprehensive performance, the multioutput response sensitivity analysis of the PMS of electric vehicle is carried out by considering the uncertainty of system parameters. Firstly, a 13degreeoffreedom analysis model of PMS is established, and the uncertain parameters of system are described by the random variables. Then, based on the summation of covariance decomposition, the first order index and the global sensitivity index of the multioutput response of system are derived. Next, a method of calculating the sensitivity indexes of multioutput response is proposed based on Monte Carlo analysis. Finally, the effectiveness of the proposed method is verified by the numerical example of the PMS of an electric vehicle. The analysis results show that the single output sensitivity analysis may not be able to accurately evaluate the comprehensive influence of parameters on the system response, and it may produce contradictory results. The proposed multioutput sensitivity analysis method can effectively evaluate the comprehensive influence of system parameters on the system response, and it can obtain more accurate sensitivity ranking for system parameters.

For the limitation of traditional type1 Smith fuzzy control in terms of inadequate time delay compensation and insufficient robustness under varying parameter driving conditions, an interval type2 Smith fuzzy time delay compensation control method is introduced for magnetorheological (MR) semiactive suspension systems. This approach incorporates the vehicle's vertical acceleration, suspension deflection, and tire dynamic displacement as the state input of the control system, enabling a comprehensive capture and response to dynamic vehicle changes. By introducing in upper and lower membership functions, the method defines clear membership intervals for fuzzy variables, which are then leveraged to calculate activation intervals under various fuzzy rules, significantly enhancing the system's antiinterference capability. Additionally, the Centerofsets algorithm is innovatively introduced into the fuzzy reduction process, avoiding redundant normalization calculation during the type reduction of type2 fuzzy sets, thereby improving the system's execution speed and realtime performance. Simulation results demonstrate that the proposed interval type2 Smith fuzzy delay compensation control strategy achieves improvement in both control effectiveness and robustness for MR semiactive suspension systems, effectively tackling complex and varied driving environment.

The damper is a core component of the suspension system, exerting significant influence on both vehicle handling and ride comfort. Traditional damper exhibits unstable response and distorted characteristics under hightemperature and highspeed conditions. Moreover, the development process heavily relies on extensive experimentation, leading to prolonged design cycles and increased cost. For this, firstly, the scheme of the continuous damping control (CDC) damper with builtin combination valve is proposedd, and the response characteristics of the valve system are quantified based on finite element method. Secondly, the nonlinear features of the damper external characteristics are analyzed, with which the hybrid model that combines piecewise models and compensation models is established to effectively capture the nonlinear gas hysteresis characteristics. Finally, all damper model parameters are identified based on measured data under different currentfrequency coupling excitation effect. Subsequently, the parameters frequencyvarying characteristics and the model accuracy are verified. The results indicate that the accuracy of the proposed hybrid model is improved by 55.91% on average compared with the piecewise model, with the error less than 10% compared with the measured data. The proposed innovative damper structure along with its characteristic modeling method can significantly enhance damper performance while simultaneously reduce development cost.

Tire matching selection is an important part of the vehicle development process. Currently, the mainstream subjective evaluation methods have problems such as poor consistency between enterprises, shortage of driver resources, and lagging evaluation nodes. In this paper, based on the subjective and objective test data of real vehicles, considering the inherent correlation between tire mechanical performance and structural parameters of vehicle handling stability, by extraction of subjective evaluation influencing factors and generation of the dimensionality reduction feature space based on correlation analysis, a subjective and objective fusion index system is established. Further, with the objective indicators and subjective rating prior trend relationship as the constraint penalty term, a tire and vehicle handling stability matching evaluation model based on subjective and objective fusion is constructed by designing an ensemble learning algorithm. The mean MSE on multi region test data is 0.247, and the predicted results show good consistency with the test results. The relevant achievements can build a quantitative evaluation system for the subjective and objective consistency of vehicle handling stability, providing support for precise selection of tire matching.

To investigate the impact of variation in tire tread depth on vehicle performance, performance tests are conducted using tires with differing tread depth as our subjects on dry, wet, and lowadhesion road surface at the proving ground. The results indicate that the braking distance decreases as tread depth diminishes on dry surface while the trend reverses compared to dry surface on wet surface, and tread depth variation has a minor impact on braking distance on lowadhesion surface. The maximum lateral acceleration and yaw rate of the vehicle initially increase and then decrease with reducing tread depth on dry surface, peaking at 4.0 mm. The maximum yaw rate follows a similar pattern to dry surface on wet surface, but there is no clear trend in maximum lateral acceleration.

With the rapid development of the electronic and electrical architecture of intelligent and connected vehicles, the demand for realtime reliability in invehicle communication networks has significantly increased. In this context, TimeSensitive Networking (TSN) has become a critical technology to meet the demand. In this paper, the implementation of the IEEE 802.1CB protocol in vehicular networks is realized, filling the gap in current research regarding the combined use of link redundancy transmission and routing planning. An innovative multipath routing strategy is proposed which balances network efficiency and reliability through dualpath transmission involving both primary and redundant paths. The core contribution of this study includes: (1) a novel NSGA2based primary path routing algorithm, which achieves the dual objectives of load balancing and low latency through intelligent path planning, and (2) an improved Dijkstrabased redundant path routing algorithm, which ensures highreliability transmission for information flows with varying priority levels. Finally, a hardwaresoftware integrated experimental framework is proposed, demonstrating that the proposed algorithms outperform existing comparison algorithms by 18.19% to 62.29% in terms of load balancing and endtoend latency, while also enhancing network reliability by 19.18% to 42.87%.

In unstructured scenes, there are often obstacles of various sizes, and the path planning process that only considers obstacle avoidance methods such as detours will lead to decrease in vehicle traffic efficiency. For these problems, in this paper an intelligent vehicle path planning method with multiple obstacleavoidance modes is proposed by integrating a layered collision detection strategy into the traditional Hybrid A* algorithm. Firstly, a doublelayer grid map is constructed based on the vehicle chassis height, and a layered collision detection strategy is designed using the body contour and fourwheel contour. Then, through a welldesigned heuristic function and cost function calculation method, the Hybrid A* algorithm can efficiently search for paths in multi obstacle scenes. Finally, the gradient descent method is used to smooth and optimize the path. Simulation and real vehicle experiment results demonstrate the effectiveness of the proposed algorithm in improving path search efficiency and significantly enhancing path smoothness. Moreover, the planned paths consider both crossing and bypassing strategies for obstacle avoidance, enabling vehicles to have better passability in multiobstacle scenarios.

In the field of intelligent connected vehicles, the recognition accuracy of incar systems for noncommand voice input in complex environment (the proportion of correct voice input recognition by the system) is of great significance. To address this challenge, in this paper a multimodal rejection model is proposed. The model is based on the opensource ChatGLM26B large language model and has undergone exclusive rejection dataset construction and model finetuning for the invehicle interaction scenario. The rejection dataset is collected from real driving scenarios, integrating voice information with the driver's facial orientation, gestures, and emotion, and other nonverbal signals to provide richer interaction information, effectively overcoming the limitation of pure language recognition mechanisms in complex environment. Through experiments, it is found that the multimodal rejection model shows higher recognition accuracy (ACC) and lower false rejection rate (FRR) on the test set compared to the pure language rejection model.