Blind equalization in wavelet domain

In this paper we propose and analyze an approach to improve the convergence rate of blind equalizers for nonstationary signals using wavelet transformation. Transform domain adaptive filters are famous for their improved convergence rate over the conventional least mean square (LMS) algorithm and also facilities for noise reduction without giving much increase in the computational cost; on the other hand, blind equalizers suffer from the poor convergence rate. We propose a wavelet domain (WD) equalization method to improve the convergence rate and discuss its advantage over other transform domain adaptive filters. This discussion is performed on two important categories of blind equalization; the stochastic gradient descent approach and cyclostationary based approach that is used in the case of blind fractionally spaced equalization (FSE) of channels.