A Blind Adaptive Matching Pursuit Algorithm for Compressed Sensing Based on Divide and Conquer

摘要

The existing reconstruction algorithms are mainly implemented based on the known sparsity of original signal, hi this paper, a new algorithm called blind adaptive matching pursuit (BAMP) is proposed in this paper, which can recover the original signal fast in the case of unknown sparsity based on the divide and conquer method. Firstly, the range of sparsity is determined by trial and error test. Secondly, the efficient support set is screened out rapidly by adaptive grouping and support extension. Last but not least, reconstructed signal is obtained by pruning. The results of simulation show that the new algorithm can reconstruct signal faster and get better precision than other similar algorithms in the same conditions.