ACTIVE FAULT-TOLERANT CONTROL FOR UNSTABLE SYSTEMS

Abstract

This paper concerns active fault-tolerant control (FTC) for unstable systems. A major obstacle to this problem is that the time delay between the fault appears and the FTC controller becomes active may destabilize the system. In this paper, an active FTC scheme based on gain-scheduled H-infinity controller and neural network is proposed. Under the assumption that the effects of faults on system matrices can be of affine parameter dependence, a reconfigurable robust H-infinity controller is developed, which is a function of the fault effect factors that can be derived on-line from the residual vector of the fault detection and isolation (FDI) mechanism. The proposed method is used for successful FTC control of a double inverted pendulum system with a fault in the tachometer loop of the motor.