An improved greedy algorithm for sparse channel estimation

Abstract

Sparse channel estimation has attracted much attention these years, especially in the area of under water acoustic communication. Compressed sensing methods are popular recently because of their efficiency and stability. In this paper, a stable and fast algorithm termed Selective Regularized Orthogonal Matching Pursuit (SROMP) is proposed based on Orthogonal Matching Pursuit (OMP). By numerical experiments, performance of this algorithm is shown in comparison to conventional LS (least square) algorithm, basic OMP and Stagewise OMP. Simulation results indicate that this methods can estimate sparse channel effectively and accurately outperforming LS and OMP.