Improved sparsity adaptive matching pursuit algorithm

摘要

For the signal reconstruction with unknown sparsity, this paper proposes an improved sparsity adaptive matching pursuit algorithm (ISAMP). Firstly, in terms of the feature that parameters k and S_t of restricted isometry property (RIP) are unknown, the proposed algorithm designs an method for the sparsity estimation. Then, the stage step-size is adaptively adjusted according to the energy ratio between the measurement vector and the reconstruction signal. Finally, after realizing the approximation of sparsity by multiple iterations, the signal is reconstructed accurately. Experimental results demonstrate that the proposed algorithm not only achieves the signal reconstruction effectively, but also obtains better reconstruction performance and lower running cost compared with similar algorithms.