Neural Network-Based Adaptive Fault-Tolerant Control for Markovian Jump Systems with Nonlinearity and Actuator Faults

The fault-tolerant control (FTC) issue is considered in this article for Markovian jump systems (MJSs) in which both nonlinearity and actuator faults exist simultaneously. The existed nonlinearity in the considered MJSs means that there exist limitations to employ the renown sliding mode control (SMC) method directly. In this work, the radial basis function (RBF) neural network (NN) technique is exploited to model the nonlinearity on which no knowledge whatsoever is available. Then, with the help of the adaptive backstepping method, an NN-based FTC approach is proposed to overcome the considered challenging case. The adverse effects, arising from the nonlinearity and the actuator faults can be completely compensated by the proposed adaptive controller. With the proposed controller and the adaptation laws, the bounded stability of the considered closed-loop plant can be guaranteed. Furthermore, only two types of adaptive parameters are adopted in the proposed approach to achieve the purpose of FTC, and this reduces the computational burden and thus extends its applicability. Finally, the effectiveness of the developed approach is demonstrated on a practical system: a wheeled mobile manipulator.