基于CS的稀疏度变步长自适应压缩采样匹配追踪算法

SPARSITY VARIABLE STEP-SIZE ADAPTIVE COMPRESSIVE SAMPLING MATCHING PURSUIT ALGORITHM BASED ON CS

通过对压缩感知(CS)理论进行研究和分析,在传统的重构算法的基础上,提出稀疏度变步长自适应压缩采样匹配追踪算法(CSVssAMP)。采用压缩采样和可变步长自适应变换的思想,解决了稀疏性未知信号的重构问题,有效地提高了数据的重构效率。相比传统的重构算法,该算法不需要预先已知稀疏度,并且每次迭代选择多个原子可以更精准地恢复低噪声信号。采用自适应变步长替换固定步长,提高了重构速率和目标精度。仿真结果表明,与现有重构算法相比,该算法的重构效果更好。

This paper studies and analyses the CS theory.Based on the traditional reconstruction algorithm,this paper proposes a sparsity variable step-size adaptive compressive sampling matching pursuit(CSVssAMP)algorithm.The idea of compressed sampling and variable step-size adaptive transformation were used to respectively solve the problem of sparse unknown signal reconstruction,which effectively improved the efficiency of data reconstruction.Compared with the traditional reconstruction algorithm,it did not need to know the sparsity in advance.Selecting multiple atoms per iteration could restore the low-noise signals more accurately.The fixed step-size was replaced by adaptive variable step-size,and the reconstruction speed and the target precision were improved.The simulation results show that the improved algorithm has better reconstruction effect,compared with the existing reconstruction algorithms.