Comparison of SPARLS and RLS algorithms for adaptive filtering

Abstract

In this paper, we overview the Low Complexity Recursive L1-Regularized Least Squares (SPARLS) algorithm proposed in [2], for the estimation of sparse signals in an adaptive filtering setting. The SPARLS algorithm is based on an Expectation-Maximization type algorithm adapted for online estimation. Simulation results for the estimation of multi-path wireless channels show that the SPARLS algorithm has significant improvement over the conventional widely-used Recursive Least Squares (RLS) algorithm, in terms of both mean squared error (MSE) and computational complexity.