Identification of nonlinear ARX model with input and output dependent coefficients

This paper proposes to identify a nonlinear system in the nonlinear ARX model from input-output data. A local linear ARX model identification is done by selecting the input and output data around the selected level. By integrating the local linear ARX models, a nonlinear ARX model with parameters nonlinearly depending on the input and output is identified. The dependence of parameters on the input and output is identified numerically and can be expressed approximately by the polynomials. The proposed method gives an approach to identify and to represent a nonlinear system which may be used to design a controller