Neural Network - based Decoupled Sliding Mode Controller Design for Discrete - time Nonlinear MIMO Systems by SPSA Algorithm

In this paper, a neural network-based sliding-mode controller design approach with decoupled method is proposed for a class of nonlinear discrete-time uncertain multi-inputmulti-output (MIMO) systems. The neural network is used to generate the proper control inputs by simultaneous perturbation stochastic approximation (SPSA) algorithm. The decoupled method simplifies the design complexity to achieve asymptotic stability for the uncertain nonlinear system with external disturbance. The proposed control scheme does not need the exactly system model to avoid the mathematical derivation. In addition, the frictional force analysis for the nonlinear inverted double pendulum system is considered to investigate the relationship between controller and frictional force. Simulation results are presented to illustrate the effectiveness of our approach.