2DOF manipulator tracking control based on fuzzy CMAC neural network dynamic inversion

A neural network based direct adaptive dynamic inversion control method is proposed for manipulator system's characteristics of nonlinear time-varying, multivariable, strong coupling. Basic control law is designed by nonlinear dynamic inversion method. Because of the nonlinear system inversion error which caused by the inaccurate model and uncertainties, fuzzy cerebella model articulation controller (fuzzy cerebella model articulation controller, FCMAC) neural network online compensation is used. Neural network make up for the shortcomings of the nonlinear dynamic inversion which need accurate mathematical model by compensating inversion error and improve the robustness of the system. The simulation results show that this method can precisely and quickly track the reference model input, and the convergence is better.