Adaptive combination of affine projection and NLMS algorithms

Abstract

We propose a novel scheme for combining two adaptation terms of affine projection algorithms with different projection orders and step-sizes. The selection of the mixing parameter that determines the performance of the proposed combination scheme is derived by the largest decrease of the mean-square deviation. The experimental results show that the proposed algorithm exhibits a fast convergence speed and a small steady-state misalignment. Moreover, the proposed algorithm achieves a low computational complexity in the steady-state. (C) 2014 Elsevier B.V. All rights reserved.