Path Replanning and Hybrid Fault Tolerant Control for Intelligent Vehicle Under Steering System Faults

Steering faults can compromise the lateral control performance of intelligent vehicles, creating a trade-off between tracking performance and degraded actuation capabilities severely. Traditional fault tolerant control methods rely heavily on accurate vehicle models, but inevitable model mismatches in real world scenarios limit their effectiveness. To address these challenges, this paper proposes a replanning based hybrid fault tolerant controller that integrates dynamic path replanning with data driven and model based fault tolerant control strategy. Specifically, a path replanning module is developed by incorporating barrier functions and path tracking objectives to refine the infeasible reference path dynamically. Furthermore, to overcome the limitations of conventional model based fault tolerant controllers and enhance the adaptability of the data driven works to unmodel part, a hybrid fault tolerant path tracking controller is designed by integrating data driven control approach into the model based approach. The model predictive controller provides the fundamental description of the system, while a data driven fault tolerant controller is constructed using online driving data to compensate simulation to reality gap. Then, the fusion strategy is employed to coordinate the different control properties and combine the advantage of both strategy. Extensive simulations and real vehicle experiments validate the superiority and effectiveness of the proposed approach in practical driving scenarios.