Fault‐tolerant control for multi‐agent systems: Addressing actuator and sensor faults in leaderless/leader‐follower architectures

Abstract

AbstractThis paper addresses the issue of actuator and sensor faults causing instability in leaderless/follower multi‐agent systems (MASs) and investigates adaptive fault‐tolerant control (FTC) algorithms to secure stability. Sensor faults can disrupt conventional observers by interfering with system state measurements. To address this issue, the descriptor approach is introduced to construct an augmented system from the original system, which converts the faults into components of the augmented system state. Next, by utilizing a sliding‐mode term to compensate for the lumped uncertainty's impact, an observer is designed to realize state estimation for the augmented system. Then, based on the estimations, adaptive FTC protocols are investigated for leaderless/leader‐follower MASs to retain stability against concurrent faults and uncertainties. To enhance control efficiency, the control gains are updated with the estimations from neighbors and proved to be ultimately upper bounded. Finally, the proposed method is validated through numerical simulations of multiple quadrotors.