Constrained Least Square based identification for Wiener systems

This paper presents a Constrained Least Square based iterative algorithm for Wiener system by using a relaxation approach based on half substitution formulation. This method formulates the problem by using half substitution and estimate the bilinear parameters of Wiener system to generate optimal estimates. This particular formulation is solved using computationally attractive relaxation by restating the problem as a constrained Least Square optimization. Simulation results are given for the Wiener model structure using polynomial and piecewise polynomial non-linearity in the presence of colored noise to show the effectiveness of the iterative method.