Self-checking Orthogonal Matching Pursuit for Compressed Sensing Signal Reconstruction

摘要

In this paper, we present a new greedy reconstruction algorithm for practical signal based on Compressed Sensing (CS), termed the Self-checking Orthogonal Matching Pursuit (ScOMP). Major innovation of ScOMP is the reconstruction process without exact value of signal sparsity or even estimated value of the optimal initial step size. The proposed algorithm improves the quality of signal reconstruction by inheriting the backtracking idea and detecting iteration with the self-checking concept. The self-checking concept makes it possible to combine with other algorithms simply by means of detecting reconstruction error in the process of signal reconstruction. The results of simulation show that ScOMP has significant advantages in the probabilities of successful reconstruction and the reconstruction accuracy, which is favorable to apply compressed sensing to practice.