An improved orthogonal matching pursuit based on randomly enhanced adaptive subspace pursuit

Greedy pursuit algorithms are widely used for sparse signal recovery from a compressed measurement system due to their low computational complexity. Combining different greedy pursuit algorithms can improve the recovery performance. In this paper an improved orthogonal matching pursuit (OMP) is proposed, in which the randomly enhanced adaptive subspace pursuit (REASP) is used to refine the estimated support set of the OMP at each iteration and hence boost the sparse signal recovery performance of the OMP. The simulation results verify the effectiveness of the proposed algorithm and show that it has good performance in noiseless and noisy cases.