Researches On Reconstruction Algorithms Of Compressive Sensing

Compressive sensing technology is an emerging information acquisition and processing technology,attracting considerable attention in many areas.Compared to traditional information collection technology,compressive sensing technology requires a smaller number of measurements,that it may be possible to break the traditional limits of sampling theory,greatly promoting the development of signal processing.As a key part of compressive sensing system,reconstruction algorithm,which directly affects the practical process of compressive sensing technology,has become the focus of industry research.In this paper,the current research status and background were introduced based on the architecture of compressive sensing,especially the researches of reconstruction algorithms.Based on the analysis on several existing matching pursuit algorithms,this paper proposed an improved algorithm.We compared the proposed algorithms with the existing algorithms by simulation experiments,where the former is shown to have an improved reconstruction performance.The main contents and innovations of this paper are as follows:(1)Introduction of three important parts of the CS theory:Low-dimensional signal models,measurement matrix and the reconstruction algorithm.Some simulation experiments on reconstruction quality and speed were given for matching pursuit,orthogonal matching pursuit(OMP),stage-wise orthogonal matching pursuit and subspace pursuit(SP).(2)By combining the smoothed LO norm algorithm and SP algorithm,an algorithm called smoothed LO norm-based subspace pursuit(SLOSP)was proposed.This algorithm uses a smooth function to approach LO norm,and gets an initial solution.Then based on this initial solution,SP algorithm is implemented to obtain the final reconstruction result.Simulation results show the improved performance of the proposed algorithm.(3)Considering the drawback of orthogonal matching pursuit,which suffers from large number of iterations,an algorithm called adaptive step orthogonal matching pursuit was proposed.By adopting the advantages of stage-wise orthogonal matching pursuit and SP algorithm,this algorithm greatly enhances the reconstruction performance.