Fault-tolerant cooperative control design for a team of car-like vehicles with actuator faults and fading channels

Abstract

This doctoral thesis proposes novel fault-tolerant cooperative control approaches for a team of car-like vehicles subject to actuator faults, fading communication channels, and external disturbances. In this study, the focus is more on the actuator faults due to the fact that actuators play a key role in the stability of physical systems and can significantly impact control systems and result in catastrophic accidents when they experience some failures. Actuator faults can result from the aging or deterioration of actuator components, leading to the actuator’s eventual bias or loss of effectiveness. This thesis aims to further investigate the impact of the actuator faults on the stability of networked control systems taking into account fading channels and external disturbances while considering problems related to the actuators that have not been addressed in the relative studies. To this end, three fault-tolerant cooperative control approaches have been developed for this purpose and validated in numerical simulations and real-time experimental tests using the latest Quanser self-driving car QCar platform. The first approach is the Distributed Fault-Tolerant Formation Control (DFTFC) which aims to control a group of car-like vehicles subject to additive actuator faults. In the design of the DFTFC, we consider the dynamics of the actual vehicles and address both the partial and severe actuator faults problem. Moreover, we provide solutions to the re-assignment formation problem for healthy vehicles in the presence of one or more vehicles with severe/complete actuator faults. Our proposed DFTFC can reduce the computation load due to the implementation of the control algorithm in a distributed manner by each vehicle and provides an overall reliable system against the sudden changes that may occur to the actuators during the formation mission. The DFTFC scheme is designed based on a High-Order Sliding Mode (HOSM) observer to detect and isolate additive actuator faults, which are superimposed on the control signals, and the super-twisting algorithm to compensate for the impact of such actuator faults. The rigorous proofs are provided to guarantee the efficacy of the proposed DFTFC in the presence of faults under investigation. The experimental results demonstrated the effectiveness of the DFTFC in providing a faster response to the occurrence of faults and good performance in handling both partial and severe actuator faults as compared to the recently reported studies. Second is the Fault-Tolerant Cooperative Control (FTCC) approach. The objective of this approach is to control a team of car-like vehicles subject to multiplicative and additive actuator faults with the presence of faded neighborhood information. The motivation for this study stems from the observation that agents exchange information over wireless networks, which inherently introduce random fading to the transmitted signals. In the cooperative control of networked systems with a restricted communication bandwidth, continuous interaction between neighboring agents would inevitably increase the network transmission load and limit the practical application of the system. It is thus important to investigate how to ensure the convergence of the team members to the desired trajectory in the presence of actuator faults using contaminated data. In the design of the FTTC, first, we investigated the impact of fading channels problem on the performance of car-like vehicles. Then, we studied the impact of the simultaneous occurrence of fading channels and actuator faults. The FTCC is designed based on the integral terminal sliding mode (ITSM) technique, which is known for its responsiveness, robustness, and stability in unknown conditions. To the best of our knowledge, non of the existing studies reported results on the actuator faults problem with faded neighborhood information. This is the first attempt to investigate the simultaneous occurrence of fading channels with actuator faults for a team of car-like vehicles. The performance of the FTTC is successfully validated in real-time experiments. Finally, in the third approach, Adaptive Fault-Tolerant Cooperative Control (AFTCC), we further investigate the fading channels problem and actuator faults with the presence of external disturbances. The AFTCC is developed based on the Non-singular Fast Terminal Sliding Mode Control (NFTSMC) to speed up the consensus tracking and convergence of the entire system. This controller is developed to handle both the multiplicative and additive actuator faults as well as the multiplicative randomness of the fading channel. The impact of faded neighborhood information and actuator faults with external disturbance on team performance is carefully analyzed. It is demonstrated by the Lyapunov stability theory that the controller can guarantee good performance of the overall system regardless of these issues. Furthermore, the synchronization of vehicles can be ensured. In comparison with existing fault-tolerant control methods, our proposed method has a simple structure and can be easily implemented in practice. Simulations are provided to validate the theoretical results.

Date	30 Sept 2023
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	Zhaoheng Liu (Supervisor) & Youmin Zhang (Co-supervisor)