一种贪婪自适应压缩感知重构

Greedy adaptive recovery algorithm for compressed sensing

为了优化压缩采样匹配追踪算法的性能,提出一种压缩采样修正匹配追踪贪婪自适应算法.该算法采用了具有理论保证的模糊阈值预选方案以避免预选时使用信号的先验信息,设置了初次裁剪门限以减少不必要的迭代,改进了裁剪方式以尽可能地提高重构精度,同时避免了裁剪阶段使用先验信息,最终实现了可压缩信号的自适应重构.仿真结果表明:在同等稀疏条件下实现了精确重构,该算法与原算法相比运算速度提高了2倍,所需观测值个数少1%,并且在稀疏度较高的情况下,该算法对噪声的抗干扰能力也优于原算法.

In order to optimize the performance of Compressive Sampling Matching Pursuit （CoSaMP）, the Compressive Sampling Modifying Matching Pursuit greedy adaptive algorithm （CoSaMMP） is proposed. Compared with the original CoSaMP, the algorithm adopts the fuzzy threshold preliminary rule with theoretical guarantee to avoid using apriori information on signals in the primary election phase, sets the initial pruning threshold to reduce unnecessary iterations, improves the pruning mode to enhance the recovery accuracy and avoid using apriori information on signals in the pruning phase, and finally realizes adaptive recovery for compressible signals. Simulation results show that for the same sparsity level, the operation speed of CoSaMMP increases by 2 fold compared with the initial algorithm, and that the required measurement number decreases about 1%, In addition, under the conditions of the high sparsity level, the algorithm have the better anti-interference ability than the initial one.