A prefiltering technique for improving the convergence of second-order Volterra equalizer

摘要

We propose a method that uses an autoregressive type prefilter for improving the convergence of the equalizer consisting of a second-order Volterra filter. In the proposed structure a reverse filter of the estimated channel is concatenated to the front part of a Volterra filter. For the conventional Volterra equalizer, when the correlation of an input signal was large, a problem occurred where the convergence of the LMS algorithm got worse. For the proposed structure, the convergence of the LMS algorithm is improved by putting the reverse filter which has the effect of reducing the correlation of the signal produced by the channel on the front part of the Volterra equalizer. We show the validity of the proposed structure by computer simulations.