Data‐driven integral terminal sliding mode fault‐tolerant control for a collection of discrete‐time nonlinear systems

Abstract

AbstractThe manuscript proposes a data‐based integral terminal sliding mode control approach to determine fault‐tolerant control (FTC) concerning a class of discrete‐time nonlinear systems (DTNS) containing sensor fault. The proposed scheme can be readily implemented because it purely utilizes the input and output data of the system to employ a model called the compact form dynamic linearization (CFDL) data model, while the one‐step forward approximator is employed to estimate the fault function term in the system. Moreover, the fault diagnosis mechanism utilizes a time‐varying threshold with prescribed performance to judge whether the fault occurs. Hence, the problem of a standard fixed threshold used at the start‐up state of the control system is resolved. Then, the discrete‐time integral terminal sliding mode fault‐tolerant control (DITSM‐FTC) method is proposed. Concurrently, the sturdiness of the proposed method is assured by hypothetical investigation. When compared with the literature, the fundamental features of the proposed method are presented as follows: (1) the problem pertinent to fault utilizing a data‐driven model is resolved; (2) a fault diagnosis mechanism with prescribed performance is proposed, and (3) one step forward approximator is utilized to devise an estimation approach of fault detection. Finally, the efficacy of the proposed method is presented by running simulations.