Event-triggered shared control for lane keeping assist system in steer-by-wire vehicles under unknown dynamics and actuator failure

This paper presents a novel shared control strategy for Human–Machine Interaction (HMI) systems, specifically targeting lane-keeping assistance by incorporating the human driver into the control loop. Specifically targeting lane-keeping assistance by incorporating the human driver into the control loop. The lateral vehicle dynamics are influenced by unknown nonlinearities such as tire uncertainties, external disturbances and potential faults in the steering actuator which pose significant challenges to stability and performance. To address these issues, a comprehensive modeling framework is developed to derive an accurate equivalent system model. A key feature of the proposed approach is a nonlinear fault-tolerant shared control system that compensates for unknown nonlinearities and actuator faults without requiring prior knowledge of the system model. This is achieved using robust approximation techniques based on adaptive Generalized Regression Neural Networks. Additionally, the paper introduces a novel authority-sharing mechanism between the human driver and the assistance controller. This mechanism use an event-triggered generator designed via Lyapunov theory, ensuring the stability of the overall closed-loop HMI system. The effectiveness of the proposed strategy is validated through both theoretical analysis and experimental evaluation, demonstrating its potential to enhance the safety, robustness and performance of shared and autonomous driving modes.