Actuator Fault Tolerant Control in Switched Systems: A Comprehensive Approach Integrating Clustering, Classification, and LMI-Based Compensation

Abstract

This paper presents a new strategy to address the challenges associated with the design of fault tolerant tolerant control (FTC) in stochastic switched systems. In fact, the problem of predicting the switched mode in discrete behavior is based on a set of input-output database. To address this, the proposed methodologies incorporate a combination of clustering and classification techniques to determine the number of sub-models. The estimated actuator fault is then compensated by controller generated for each sub model. The controller gains are given by solving the Linear Matrix Inequality (LMIs), which is derived based on the general Lyapunov function. The fast convergence of the actuator fault, which is estimated using the proposed data-driven method, minimizes the instability time of the switched system after fault occurrence and attenuates the disturbance. Finally, the effectiveness of the proposed method against actuator faults is validated through two numerical analyses. The first example shows a faulty stochastic switched system with two modes and the second example shows a practical application of a double cart system with elastic coupling.