An Adaptive Control with RBF Neural Network for Discrete-Time Nonlinear System

Discrete-time nonlinear systems are a prevalent processes that operate on sampled data rather than continuous signals, play a pivotal role in modern engineering and technology. Due to the high nonlinearities, how to efficiently control the discrete-time nonlinear systems is a challenging problem. In this paper, we propose an adaptive control with RBF neural network (AC-RBF). In the proposed AC-RBF, RBF neural network (RBFNN) is first designed to adaptively approximate the controlled object by its dynamic weight adjustment mechanism. Combined with the approximation information, the output of RBFNN can be considered as the control law function, among which the weight parameters are the independent variables. By solving the control law function based on the input data and initial information, we can obtain the optimal control law sequence with a certain sampling interval. Finally, the proposed AC-RBF is used to control the discrete-time nonlinear system and the simulation results demonstrate its effectiveness.