A Novel Sparse Signal Reconstruction Algorithm Based on $l_{0}$ Minimization

Abstract

Greedy reconstruction algorithms have fast signal reconstruction speed and low computational complexity. They employ the local optimization strategy so that they tend to generate sub-optimal solutions. To solve the sub-optimal solution problem, we suggest a novel sparse signal reconstruction algorithm, called genetic sparse adaptive matching pursuit algorithm (GSAMP). The algorithm includes two main steps: First, by setting different step size, the sparse adaptive matching pursuit (SAMP) algorithm can obtained different solution as the initial population of the genetic algorithm. The genetic algorithm is designed in the second step to obtain the optimum solution. We design three groups of experimental to evaluate the performance of GSAMP. Experimental results show the proposed algorithm has more excellent performance than some classical reconstruction algorithms.