Neural Network Based Decoupling Control for a Class of MIMO Nonlinear Systems

摘要

A hybrid decoupling controller, consisting of conventional PID controllers and neuro-controllers, is proposed for a class of multi-input-multi-output (MIMO) systems. This controller combines the learning capabilities of artificial neural networks and the global asymptotic stability of the conventional PID controllers. It provides a better control for coupling nonlinear systems which otherwise would have been difficult to control through conventional PID controllers.