Output Stability Control Design for a Class of Nonlinear Systems based on Nested BP Neural Network

In this paper, the concept of the state domain of the hidden layer neuron is introduced for the three-layer BP (Back Propagation) neural network. By employing the state domain and the partition of unity method to modularize the hidden layer, and a new artificial network called Nested BP neural network is constructed to reduce the computational complexity and improve the learning quality. The Nested BP network is employed to approximate the unknown nonlinear items in controlled systems. Compared to the usual BP network, the control effect in this paper is greatly improved. Finally, an example shows that the proposed method is superior to the usual BP neural network in learning quality and control effect.