Performance Analysis of Regularized Adaptive Filter for an Acoustic Echo Cancellation Application

Currently, adaptive echo cancellation algorithm is mainstream technology in the field of noise cancellation. In general an adaptive filter performs more efficient than normal filters, but suffers with ill-posed problem. This problem is due to echo in the observation data or solving linear system of equations. By using regularization concept we optimize this problem. In this paper we proposed a regularization parameter for four important adaptive algorithms the normalized least-mean-square (NLMS). Simulations performed on an ANC application, which is basically a system identification problem, with different ENRs