An improved reconstruction algorithm for non-Gaussian signal in compressive sensing

The belief propagation-based reconstruction algorithms with the sparse sensing matrix Φ are quite robust against measurement noise and have fast convergence speed. However the reconstruction performance of these existing belief propagation-based reconstruction algorithms is not so satisfactory for non-Gaussian signal. To improve the reconstruction performance further we consider an improved reconstruction algorithm with iterative support detection and pseudo-inverse decoder. Simulation results show that the improved algorithm outperforms the normal belief propagation-base algorithms for non-Gaussian signal.