An LMI-based convex fault tolerant control of nonlinear descriptor systems via unknown input observers

This paper proposes a fault tolerant control scheme for nonlinear systems in descriptor form.The approach is based on the design of an unknown input observer in order to estimate the missing state variables as well as actuator faults, such design is carried out once a proper estimation error system is obtained via a recent factorization method; then, the estimated signals are employed in the control law in order to drive the states asymptotically to the origin despite actuator faults.The designing conditions are given in terms of linear matrix inequalities.Numerical as well as physical systems are used to illustrate the advantages of the proposal.