Robust fault tolerant control based on uncertainty and disturbance estimation

Abstract

Fault tolerant control (FTC) becomes essential to achieve satisfactory performance in the occurrence of unexpected scenarios such as faults and failures. Therefore, being a cost effective strategy to achieve increased reliability in automatic systems, the FTC has been an active research area. In the literature, many robust control strategies exist to cope up with system uncertainties and unmeasurable external disturbances. However, the designs usually do not address the issue of fault tolerance. Since practical systems are also prone to faults apart from system uncertainties and disturbances, a design that simultaneously takes care of these issues is needed. In essence, design of FTC which is robust in the presence of system uncertainties and disturbances while fault occurs becomes highly desirable. Addressing this aspect, this research work proposes a robust FTC design. newlineAt the outset, robust control strategies using the disturbance estimation and compensation approach are studied. From the many available approaches, an Uncertainty and Disturbance Estimator (UDE) approach is used to robustify the Nonlinear Dynamic Inversion (NDI) based FTC scheme. The UDE technique provides estimate of the total disturbance enabling its rejection and thereby achieving robustness to the NDI controller. Since the NDI design needs availability of successive derivatives of output, the same are obtained through an UDE robustified observer making the NDI control design implementable. Lyapunov stability theory is used to establish stability for the proposed UDE based controller observer structure. The efficacy of the proposed control scheme is validated through application to quadrotor and aircraft control problems. Monte Carlo simulations and experimental validation on Quanser s 3 DOF Hover setup have also been carried out to bring out effectiveness and feasibility of the proposed design. It is shown that unlike in many FTC designs, the proposed design offers robustness against system nonlinearities, parametric variations,