Estimation of Wiener Model Based on Neural Fuzzy Network

This paper proposes a two-stage parameter estimation approach of Wiener model based on correlation analysis method and Taylor series expansion theory. The developed Wiener model is characterized through the dynamic block modeled by a rational transfer function, followed by a nonlinear block based on neural fuzzy network. The input test signal consisting of Gaussian signal and random signal is applied to the parameter separation and estimation of Wiener model. Firstly, based on the input-output data of Gaussian signal measured, correlation analysis is used to obtain the parameters of the linear block. Then, using Taylor series expansion theory and clustering algorithm, the nonlinear block parameters are estimated based on random signals. Through theoretical derivation and experimental results, it can be seen that this method can usefully estimate the Wiener model with output noise and obtain favorable estimation accuracy.