Adaptive Neural Fault-Tolerant Control for a Class of Stochastic Switched Nonlinear Systems

This paper addresses the adaptive neural fault-tolerant control (FTC) problem for a class of stochastic switched nonstrict-feedback nonlinear systems, which have actuator faults that incorporate loss of effectiveness, stuck, and outage. Based on the character of the Gaussian function, the problem of nonstrict-feedback form is solved well. In the process of designing the controller, neural networks (NNs) are utilized to estimate the unknown functions. The problem of actuator faults is handled by designing the FTC method that is obtained by introducing a smooth function and backstepping technique. Then, the control effect satisfies that all the signals in the resulting closed-loop system are bounded and the tracking error converges to a small neighborhood around the origin. To illustrate the high efficiency of the proposed control method, a vivid simulation example is given in the end.