Sparse Adaptive Channel Estimation based on l0-PRLS Algorithm for Underwater Acoustic Communications

Abstract

Underwater acoustic (UWA) channels usually manifest time-varying and sparse properties, motivating channel estimations based on sparse adaptive filtering algorithms. Existing sparse adaptive filtering methods are designed by applying a sparse norm regularization, employing a proportionate updating step size, or combining above two operations. Currently, the least mean squares (LMS) type sparse adaptive filtering algorithms have been extensively investigated for the estimation of UWA channels. In contrast, the recursive least squares (RLS) type methods are much less studied. We recently proposed a RLS-type sparse adaptive filtering algorithm named l <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</inf> -PRLS, which showed superiority for UWA channel estimation. In this paper, we substitute the l <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</inf> norm in l <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</inf> -PRLS by an approximation of l <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</inf> norm, leading to the l <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</inf> -PRLS algorithm. The l <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</inf> -PRLS was then applied for sparse channel estimations. Simulation and experimental results are provided to demonstrate its advantage over the standard RLS and its other sparse variants.