Active sensor faults tolerant control for Takagi-Sugeno multiple models

摘要

In this communication, a method allowing the conception of a sensor faults tolerant control for the case of non linear systems described by Takagi-Sugeno models. The faults are initially estimated using a proportional integral observer. The synthetized control depends on the estimated fault and the error between the state of the system of reference and the state of the system affected by faults. The conditions of the observer convergence and of the control existence are formulated in terms of Linear Matrix Inequalities (LMI). The formulation in LMI shows that the synthesis of the control and the observer can be independently made. The theoretical results are validated on a noisy non linear system affected by sensor faults.