Sparse Signal Reconstruction Algorithm Based On Residual Descent

To solve the problem of a compression sampling matching pursuit algorithm (CoSaMP) poor reconstruction result and inaccurate index selection, an improved reconstruction algorithm, called residual descent OMP (RdOMP), is proposed. The proposed algorithm introduces a residual comparison strategy to enhance the effectiveness of the backtracking indexes. This backtracking strategy, which is based on the residual descent, can flexibly select backtracking indexes. Without increasing the computational complexity, the proposed algorithm has a higher exact reconstruction rate. Different simulations show that the proposed algorithm can restore the sparse signal efficiently and reconstruct the sparse signal with a high probability.