Research On Sparse Signal Compression Perception Reconstruction Algorithm

Traditional Nyquist sampling requirements sampling frequency is more than twice the signal bandwidth of sparse multi-band signal samples will cause the sampling frequency is high, a large amount of data, waste transport and storage resources. From the analysis of compressed sensing signal structure, the use of sparse characteristic signals at the sampling rate is much less than the Nyquist sampling rate conditions, with a random sample capturing discrete samples of the signal, and then through a nonlinear signal recovery algorithms reconstruction. Compressed sensing signal sampling and compression at the same time, can effectively reduce the sampling frequency and the amount of data and reduce information redundancy, saving storage space.Compressed sensing core theory (1) signal sparse representation(2) structure measurement matrix; (3) reconstruction algorithm design. Compressed sensing sparse representation is the premise, and select the measurement matrix structure is conducive to reconstruction algorithm design and implementation; reconstruction algorithm design is the key part. Sparse representation, measurement matrix construction, reconstruction algorithm design, the three interlocking are indispensable.This paper studies the compressed sensing sparse signal reconstruction algorithm. Firstly, a brief review of the theory of compressed sensing, introduces sparse representation of the signal, the measurement matrix construction and reconstruction algorithm. Then, introduce common Sub-Nyquist sampling system that is set to Multi-Coset sampling (MC), Random Demodulator (RD), Modulated Wideband Converter (MWC), and highlight the MWC; Thesis focuses on compressed sensing recovery algorithm, and Orthogonal Matching Pursuit (OMP) algorithm, Regularization Orthogonal Matching Pursuit (ROMP) algorithm, Compressed Samples Matching Pursuit (CoSaMP) algorithm, and the OMP algorithm is proposed based on a new algorithm-Improved Orthogonal Matching Pursuit (IOMP) algorithm. This paper gives the algorithm steps above four algorithms, flow charts and the like. Using Matlab simulation platform above algorithm simulation analysis, experimental results show good performance thesis proposed IOMP reconstruction algorithm, to achieve accurate reconstruction.