Robust fault detection for uncertain switched systems

In this paper, the problems of fault detection and estimation (FDE) are addressed for a class of discrete time switched systems subject to bounded uncertainties and unknown input disturbances. Two design methods for robust fault estimation using switched filters are proposed. These problems are addressed by minimizing the L-2-gain from unknown input and uncertainties in the so called switched H-infinity problem. Stability analysis is performed by using switched Lyapunov functions and formulated by linear matrix inequalities (LMI). In the first method a functional filter for fault detection is proposed. For the second one, the fault estimation is performed by adding loop shaping and matching filters in the design procedure. Multiple Lyapunov function is proposed to proof the convergence of the filter. A family of solutions is proposed for filter reconstruction after linearization steps using the elimination lemma, offering one more degree of freedom in the design and less number of LMIs. Numerical example is given to illustrate the effectiveness of the proposed designs. (C) 2017, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.