Using neural network realize the identification of nonlinear system

System identification is a theory and technique which use input and output data to build mathematical model and provide a mature method for linear identification systems at present, but for the nonlinear identification system we have not yet found easy unity mathematical methods.Neural network has strong self-learning ability and the nonlinear mapping ability thus it has great potentialities in nonlinear identification system. In this paper we discuss the theory and characteristics of neural network for the nonlinear system, and the two typical structure of neural network identification system.