Convergence of CMAC network learning control for a class of nonlinear dynamic system

Abstract

Due to the high nonlinear property, the convergence of neural network based closed-loop system are very difficult to analyze. In the paper, we adopt cerebellar model articulation controller (CMAC) as an iterative learning controller, the convergence of the closed-loop system is analyzed based on the iterative learning control theory. Simulation results verify the analysis.