This paper presents a new approach to Fault-Tolerant Control (FTC), addressing the complexities inherent in vehicle dynamics and steering actuator faults. By combining the sliding mode (SM) method and the Generalized Regression Neural Network (GRNN) for their compensation, this novel approach introduces a single estimator to consider a complete dynamic system that incorporates all nonlinearities and actuator faults. Inspired by conventional Model-Free Control (MFC) techniques, which often rely on specific model assumptions, we propose a new adaptive model-free control structure based on local measurements. Furthermore, we introduce an innovative event-triggered control mechanism that activates control actions only when necessary, where the system closed-loop stability is demonstrated through a rigorous Lyapunov theory analysis. This strategy not only enhances system efficiency but also reduces energy consumption and processing overhead in computing systems. The effectiveness of our approach is demonstrated through numerical simulations and experimental tests, particularly in lateral Lane-Keeping Assist (LKA) systems for Steer-by-Wire (SBW) vehicle. Overall, our contributions advance the field of fault-tolerant control, offering practical solutions for real-world applications.
Event-Triggered Adaptive Fault-Tolerant Control Based on Sliding Mode/Neural Network for Lane Keeping Assistance Systems in Steer-by-Wire Vehicles
IEEE Transactions on Intelligent Vehicles ; 10 , 3 ; 1597-1608
01.03.2025
3438809 byte
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
SAGE Publications | 2021
|