Robust fault-tolerant control of networked control systems with stochastic actuator failure

Fault-tolerant control of networked control systems (NCSs) with random actuator failure is studied. An innovative model is presented for this problem. It includes three sources of uncertainties, namely uncertainties in the plant model, uncertainties in networked communications and uncertainties in possible actuator failure/malfunction. Other main features are: (i) the fault statistics of each actuator is individually quantified, and (ii) a united framework is proposed to have logic zero-order-holders embedded in the NCS. The latter enables actuators - when in normal operation - to use the latest actuating signals available to them. Based on the Lyapunov-Krasovskii functional, three theorems are proved in the study for the system stability and controller design. Theorem 1 gives a matrix inequality for the system asymptotical stability in the mean-square and is the foundation of the other two theorems. Theorem 2 shows a stability condition regarding the design of a robust state-feedback control for the system under study. Finally, Theorem 3 gives a modified stability condition that can be employed for actual design. A numerical example is presented to show how such a robust controller can be designed.