Active Fault-Tolerant Predictive Control of Networked Systems with Time-Varying Actuator Faults

This paper proposes a networked active fault-tolerant predictive control strategy for networked systems subject to time-varying actuator faults and random communication constraints. An adaptive Kalman filter is adopted to estimate the system states and actuator faults simultaneously, based on which, a fault-tolerant predictive controller is developed to reduce the system output tracking error caused by the time-varying actuator faults and random communication constraints. The closed-loop system is a bidirectional time-lag bounded stochastic switching system. The stability conditions of the system are derived by using the switching Lyapunov method. Finally, the proposed method is applied to a DC motor system and its effectiveness is verified.