Optimal iterative control of unknown nonlinear systems using neural networks

An optimal iterative controller of unknown nonlinear systems using neural networks is presented. Initially the multilayer feedforward neural network is trained to learn the forward dynamics of the plant off-line. Then the control law is derived by the optimal iterative method based on the neural network during each sampling interval. The optimal iterative controller predicts the feedforward control action, while the fixed feedback compensating controller actuates the feedback stabilising signal. Finally, simulation results show that the performance of the controller based on the proposed scheme is efficient.