The Smith neural networks pre-estimated control for a kind of system with large time-delay

This paper presents an improved Smith-NN pre-estimated control method which is designed to overcome the effect caused by controlled object´s time-varying, nonlinear and uncertainties when large time-delay is compensated. The studying of the closed loop neural controller and the Smith pre-estimated controller designing and testing prove the method is feasible. From control system simulation, especially experiments on the pH process and dynamic ration weighing process, confirm the effectiveness of the method to overcome large time-delay. This method can easily be realized by intelligent tools based on the assembly language