Integrated Fault Estimation and Fault-Tolerant Control for Switched Stochastic Systems With Unknown Input

Abstract

In this paper, fault estimation (FE) and fault-tolerant control (FTC) are studied for switched stochastic nonlinear systems with unknown input and faults. The dynamic unknown input observer (DUIO) and disturbance observer are designed, and the system state, faults, and disturbances can be reconstructed. The DUIO designed in this paper is the extension of the traditional unknown input observer (UIO) and the proportional-integral observer. The unknown input, which contains unmodeled disturbance, uncertainty, can be decoupled by the estimation error system. A dynamic output feedback controller based on observers is designed to compensate for actuator fault and modelable disturbance. The integration scheme of FE and FTC is designed, and the one-step linear matrix inequalities (LMIs) conditions are constructed, which avoids the coupling of observer and controller parameters. The root-mean-square (RMS) gain is used to attenuate the unknown input and unmodelable disturbances in the closed-loop systems, which can remove some common assumptions. Finally, the feasibility of the proposed method is demonstrated through simulations.