Robust State Estimation And Fault-tolerant Control With A New Sliding Mode State Observer

Modern control systems are getting more and more complicated. To guarantee thestability of the plant and maintain a desired productivity, fault detection and fault-tolerantcontrol methods are being more and more concerned. Usually, fault diagnosis is imple-mented by constructing residual signal from the output of the plant and the output ofobserver, then the residual signal is used to determine whether there is fault and where itoccurs. Recent years, fault reconstruction is becoming a powerful alternative to the resid-ual method, and is being widely studied by scholars. Fault reconstruction method can notonly diagnosis and separate fault, but also reconstruct the fault and give the estimation ofit’s value.This paper summarizes the results of previous research, and constructs a new kindof sliding mode observer based on PID control and sliding mode controller for to imple-ment fault reconstruct for complicated control systems. This paper is constructed by thefollowing parts.1. The first part designed a robust sliding mode observer for systems with Lipschitznon-linearity, and designed a method to construct a generalized augmented system, whichhave both the state variable of the original system and the disturbance of fault in it’saugmented state variable. Then we show how to designed the feedback control scheme,and give the condition to guarantee the stability of the overall close-loop system and theerror system.2. The second part popularize the method from the first part to a Markovian-jumpsystem. Then, we studied the stochastic stability of the Markovian-jump system by usingthe week infinitesimal operator.3. Based on the first two parts, we designed a fault reconstruction and robust-tolerance control method for the Markovian-jump system.