Bo Leng

Accurate estimation of the tire-road friction coefficient (TRFC) is crucial for maximizing the performance of the actuators in the intelligent electric vehicle chassis. The uncertainty quantification of TRFC estimation is rarely reported in existing studies. In this article, we propose an uncertainty-aware TRFC estimation framework based on evidence theory to simultaneously output point estimates and confidence intervals for TRFC. First, approximation modeling errors in the tire model fitting process are thoroughly analyzed and systematically categorized into two types: random errors caused by measurement errors in tire test data and systematic errors resulting from tire model simplification. Then, the propagation characteristics of tire model uncertainty are analyzed using Dempster–Shafer evidence theory to quantify disturbances from random errors in the estimator. For coupled excitation conditions, excitation level factors based on the friction circle are introduced to construct an uncertainty-aware dynamics fusion estimator. The input-to-state stability of the designed fusion estimator is guaranteed using Lyapunov functions. The proposed framework was validated on real-world proving grounds using both A-segment and D-segment electric vehicles. Qualitative and quantitative comparisons with unscented Kalman filter and dropout-based methods demonstrate that the proposed framework achieves superior point estimation accuracy and reliable confidence intervals under complex coupled excitations.

Reinforcement learning (RL) has shown excellent performance in solving decision-making and control problems of autonomous driving (AD), which is increasingly applied in diverse driving scenarios. However, driving is a multiattribute problem, leading to challenges in achieving multiobjective compatibility for current RL methods, especially in both policy updating and policy execution. On the one hand, a single value evaluation network limits the policy updating in complex scenarios with coupled driving objectives. On the other hand, the common single-type action space structure limits driving flexibility or results in large behavior fluctuations during policy execution. To this end, we propose a multiobjective ensemble-critic (MoEC) RL method with a hybrid parametrized action for multiobjective compatible AD. Specifically, an advanced MORL architecture is constructed, in which the ensemble-critic focuses on different objectives through independent reward functions. The architecture integrates a hybrid parameterized action space structure, and the generated driving actions contain both abstract guidance that matches the hybrid road modality and concrete control commands. In addition, an uncertainty-based exploration mechanism that supports hybrid actions is developed to learn multiobjective compatible policies more quickly. The experimental results demonstrate that, in both simulator-based and HighD dataset-based multilane highway scenarios, our method efficiently learns multiobjective compatible AD with respect to efficiency, action consistency, and safety.

Deep Reinforcement Learning (DRL) is becoming a prominent method for autonomous driving due to its strong capability to generate complex driving policy. However, DRL motion planning still has limitations in safety performance including learning quality, convergence speed and the safety guarantee. To this end, this work proposes a safety-enhanced deep reinforcement learning method with dynamic safety guidance (DSG-DRL) for lane-change motion planning. It bears the following key features: 1) Able to learn a safer DRL driving policy by additionally including potentially unsafe behaviors; 2) Able to accelerate learning a safe policy by making dangerous driving experiences impressive; 3) Able to further enhance the driving safety by avoiding unexpected reckless action. The proposed DSG-DRL motion planner dares to make mistakes to learn the safe driving policy better and faster. By evaluating anticipated risk, it learns not only from the maneuvers right at the moments of collisions, but also from the dangerous maneuvers leading towards collisions. Besides, risk driving experiences are enhanced with additional memory batches and sampling prioritization. Moreover, reckless actions can be prevented by dynamic constraints both in training and testing, which further improves the safety performance. Simulation validation shows that the proposed method can learn a safer driving policy with faster convergence speed, achieving the high safety performance while keeping the driving efficiency.

Stability constraints are crucial in the trajectory tracking control of autonomous electric trucks to ensure driving safety. However, varying road conditions and the nonlinear characteristics of the vehicle dynamics system affect the accuracy of stability constraints. A stability-oriented control strategy (SOCS), which accurately represents stability constraints through differential manifold safety boundaries, achieving coupled control of trajectory tracking and stability, is proposed in this paper. Firstly, to accurately represent stability constraints, the nonlinear characteristics of the vehicle dynamics system, the coupling effects across different dimensions, road conditions, and control inputs are thoroughly incorporated. The safety boundaries of the longitudinal, lateral, and vertical nonlinear vehicle dynamics system are constructed using differential manifold theory, and the proposed improved inverse trajectory manifold calculation method (IITM) is applied to accurately solve the safety boundaries. Secondly, based on the constructed safety boundaries, the control objective priorities and stability control weights are dynamically adjusted in SOCS, thereby improving the tracking accuracy and stability of autonomous electric trucks under complex driving conditions. Finally, the effectiveness and feasibility of SOCS are validated through simulations and hardware-in-the-loop (HIL) testing.

Lane-change trajectory planning is a key technology for ensuring the driving safety of autonomous electric trucks. However, under extreme driving conditions such as high speeds or low tire-road friction coefficient, ensuring both the safety of trajectory planning and the stability of vehicle motion remains a major challenge. This paper proposes a lane-change trajectory planning strategy based on dynamic safety boundaries (LCTP DSB), aiming to achieve coordinated optimization among trajectory smoothness, collision avoidance, and vehicle stability. First, differential manifold theory is introduced to characterize the evolution of safety boundaries in nonlinear systems. The impacts of external disturbances and control inputs on vehicle stability are analyzed, and dynamic constraints for lane-change trajectory planning in autonomous electric trucks are established. Second, to efficiently solve the dynamic constraints, a sequential region filtering and progressive optimization (SRFPO) method is proposed. By partitioning the safety boundary, this method enables rapid screening of feasible solutions and global optimal solution search. In the trajectory planning process, the transfer position is dynamically selected in response to the current dynamic state of the vehicle. The trajectory is divided into a lane-change phase and lane-alignment phase, and the final optimal lane-change trajectory is obtained through trajectory stitching and optimization. Finally, the proposed method is validated via simulation and HIL testing, confirming its effectiveness in exploiting vehicle stability limits during trajectory planning.