A New Method of Error Reduction for Affine Projection Algorithm

Abstract

This paper proposes a new method of reducing the steady-state error for the Affine Projection Algorithm (APA) using Error Reduction Factor (ERF) to yield Error Reduction Affine Projection Algorithm (ER-APA). The proposed method is very simple but highly effective. In the paper, we analyze the ER-APA for calculating theoretical filter convergence. Through experiments, it is shown that the ER-APA sufficiently reduces the steady-state error for any value of the dimension of projection, while preserving the fast convergence property of the APA.