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.
