Adaptive Matching Pursuit Method Based on Auxiliary Residual for Sparse Signal Recovery

Abstract

Greedy pursuit methods are widely used for compressive sensing (CS) and sparse signal recovery due to their low computational complexity. In this paper an adaptive matching pursuit is proposed, which is based on the backtracking-based adaptive orthogonal matching pursuit (BAOMP) and uses auxiliary residual to make correlation test to add more correct atoms per iteration. The proposed method can be regarded as an improved BAOMP. The simulation results show that it has better performance to those of some other greedy pursuit methods. Finally the experiment of CS-based ISAR imaging verifies the effectiveness of the proposed method.