Detecting sensor fault for nonlinear systems in T–S form under sampled-data measurement: Exact direct discrete-time design approach

Abstract

A direct discrete-time design methodology for sampled-data sensor fault detection for nonlinear systems in Takagi–Sugenos form is proposed. Contrary to the conventional schemes in this way that rely on an approximate discrete-time model of the nonlinear system, our result is established based on an exact one. Condition to design the observer and the residual gain under an H-/H∞ performance criterion is presented in matrix inequality format. An example is given to illustrate the effectiveness of the proposed methodology.