MODEL REFERENCE NEURAL NETWORK CONTROL FOR NONLINEAR SYSTEM

In this paper, a new neural network control strategy is proposed for unknown nonlinear dynamical systems, using dynamical multilayered BP network to identify and still network to control.The controller has simple substruct, better convergence and stronger extensive ability.Simulation shows the control strategy is effective for several typical nonlinear systems and adaptive for multivalued counter reflection systems, and can be used in realtime online control.