Identification of Nonlinear Systems Using Recurrent Neural Networks
Ren Xuemei · Control theory & applications · 2001
This paper proposes a new type of recurrent neural network for the identification of a class of unknown nonlinear system. It is proved that the proposed network with appropriate conditions can represent unknown input_output relationship of nonlinear systems. The dynamic backpropagation algorithm is employed to estimate the weights of both the feedforward and feedback connections in the networks. The proposed schemes have been successfully applied to modeling nonlinear plants.