Low rank transform domain adaptive filtering

Abstract
This paper introduces a least squares, matrix based frameworkfor adaptive filtering that includes Normalized LMS,Affine Projection, and Recursive Least Squares as specialcases. We then introduce other transform domain basedmethods and show how to create optimal low rank versions.We also discuss efficient implementation of our method andits use in an echo canceller.

This publication has 4 references indexed in Scilit: