Nonlinear Multivariable Decoupling PID Control Using Neural Networks

A nonlinear multivariable decoupling PID controller is derived from generalized minimum variance control law, which consists of a PID controller with decoupling design and a feedforward compensator for the unmodeled dynamics. Then a nonlinear multivariable decoupling PID control algorithm using neural networks is proposed. By using neural networks online estimating and compensating the unmodeled dynamics, the proposed algorithm has the adaptive capability to the variations of both parameters and structure of the process. Under some assumptions, it is proved that all the signals in the closed-loop system are globally bounded and the tracking error can be made less than any specified constant over a compact set by properly choosing the structures parameters of neural networks. Simulation results show the effectiveness and robustness of the proposed algorithm