基于压缩感知SAMP算法的改进

The Improvement of SAMP Algorithm Based on Compressive Sensing

压缩感知是近年来兴起的以较大压缩比率并且能够以较好的效果恢复原始信号的一种新型理论。本文针对SAMP算法在重构精度相对较差以及运算时间较长的问题,提出一种新的SAMP改进算法。该算法首先将稀疏信号进行分块后再重构,其次在重构算法过程中开始对信号稀疏度进行初始估计,然后采取变步长增加稀疏度的方法逐步求出原始信号的最优估计值。理论分析和实验表明,改进后的算法减少运算时间,提高重构精度,增进实际使用价值。

Compressive sensing is new theory that can exactly recover the original signal under the condition that the signal is compressed by the larger compression ratio. In this paper,an improved algorithm is proposed for enhancing the reconstruction precision of Sparsity Adaptive Matching Pursuit（ SAMP） and reducing the operation time. Firstly,the improved algorithm reconstructs original signal by the sparse block signals,then reduces the time by initial estimation of the sparse signal and approaches the real sparse level by variable step length. The analytical theory and simulation results show that not only significant reconstruction performance improvement is achieved,but also the operation time is much more faster than the original SAMP algorithm.