Over Adapted LMS Algorithm with Codified Error

Abstract

In this paper a modification to the mean square algorithm (LMS) appears, with the objective to reduce the floating point operations. This reduction allows to obtain a simple implementation of a digital adaptive filter, inside an integrated circuit, because a reduction of the floating point operations is made. The obtained advantages with this modification are the over adaptation of LMS algorithm, in order to increase the speed of convergence or higher frequency signals processing