An improved proportionate normalized least mean square algorithm for sparse impulse response identification

Abstract

In this paper after analyzing the adaptation process of the proportionate normalized least mean square (PNLMS) algorithm, a statistical model is obtained to describe the convergence process of each adaptive filter coefficient. Inspired by this result, a modified PNLMS algorithm based on precise magnitude estimate is proposed. The simulation results indicate that in contrast to the traditional PNLMS algorithm, the proposed algorithm achieves faster convergence speed in the initial convergence state and lower misalignment in the stead stage with much less computational complexity.