Blind equalization using Bayesian technique and wavelet networks

A new blind equalization algorithm based on suboptimum Bayesian symbol detector and discrete wavelet network is presented. First, the shortcomings of well-known Bayesian blind equalization is discussed, then the method is modified in order to be used for nonlinear channels. It uses discrete wavelet networks instead of linear models to cover the nonlinearities and guarantee global optimization. It is shown that the proposed method is applicable for elimination of non-symmetrical nonlinear distortions caused by even harmonies.