Activator-Based Economical Distributed Fault-Tolerant Control Against Possible Actuator Outages With Guaranteed Performance

Abstract

This paper concerns a distributed control strategy to deal with possible actuator outages of multi-agent systems (MASs). With the designed strategy, the control process is monitored by a fault detection system (FDS). Once the actuator of one agent in use partially or completely fails, the FDS will detect the fault timely and a controller activator will produce its effect. By this means, the healthy actuator can replace the faulty one for control at the delicately designed instants, and the prescribed performance of the MAS can always be satisfied in the presence of possible actuator outages with the help of prescribed performance functions. Moreover, to achieve more diversified constraints to meet wider requirements in practice, the proposed strategy is further extended to combine with a time-varying barrier Lyapunov function so that the time-varying output constraint can be achieved, which makes the designed fault-tolerant control more flexible and sensitive in dealing with failures. With the proposed distributed control method, uncertain actuator failures, especially the actuator outage, can be addressed in the presence of disturbances and uncertain dynamics. Since the failure is allowed to occur multiple times with only one actuator operating for control at any time, the designed approach can be more economical in energy saving compared with the existing methods. Numerical simulations are provided for multiple agents to verify the effectiveness of the proposed algorithm. Note to Practitioners—In this paper, a control strategy is proposed to handle the actuator failure of a class of MASs. To deal with possible actuator outages, which pose a threat to system safety, actuator redundancy is introduced in the proposed strategy. By developing a controller activator to specify the activation time of healthy actuators with the help of a designed FDS, the prescribed performance of the agents can always be guaranteed and the tracking error can be constrained within specified time-varying function curves as expected. Since the proposed strategy enables the successive triggering of actuators, only one actuator works at any instant, which contributes to its energy-saving efficiency. Since the proposed strategy can accommodate possible actuator outages in an economical way with guaranteed prescribed performance, it can be used in various industrial scenarios where high safety of agents is required, for instance, multiple aircrafts, multiple high-speed trains, and multiple industrial robot systems.