A Fully Actuated System Approach to Fault Detection Set-Valued Observer Design for Discrete-Time Nonlinear Systems

This paper investigates fault detection set-valued observer design for discrete-time nonlinear systems. Instead of employing linearized models (e.g., linear parameter-varying models and Takagi-Sugeno fuzzy models), we adopt the fully actuated system approach which enables us to effectively cancel nonlinearity and obtain linear system dynamics. Consequently, the dual objectives of stabilizing the closed-loop system and designing fault detection set-valued observers are achieved concurrently. Furthermore, the parameters of the controller and set-valued observer are determined through the solution of the generalized Sylvester equation. Finally, the efficacy of the proposed method is illustrated via a simulation example.