Optimized Algorithm Based on Random Filter for Measurement Matrices
ZHENG Yan1,LI Jian1,LI Zhi1,ZHANG Jun-peng2
1(College of Electronics and Information Engineering,Sichuan University,Chengdu 610065,China)
2 (Luoyang Ruichang Petrochemical Equipment co.,Ltd.,Luoyang 471000,China)
Abstract:Random measurement matrices often need significant space requirement for storage and high computational cost for signal reconstruction in compressed sensing.Hence,a optimized algorithm based on random filter is proposed to solve these problems of random measurement matrices.Random filter is a circulant matrix,where the nonzero entries are drawn from function of chebyshev window,hamming window,flat top weighted window or blackman window.Then,every row of this circulant matrix is orthonormalization.The numerical experiments show it outperforms random filters and chaos filters while being possible to exactly reconstruct original sparse signals in the time and frequency.what′s more,reconstruction performance of hamming window optimized matrix for sparse signals in the time domain is the best than others.Reconstruction performance of flat top weighted window or blackman window optimized matrix for sparse signals in the frequency domain is the best than others.Reconstruction performance of chebyshev window optimized matrix for sparse signals in the time domain and in the frequency domain is the best than others.
郑岩,李健,李智,张俊朋. 一种基于随机滤波器的测量矩阵优化算法[J]. 小型微型计算机系统, 2016, 37(3): 632-636.
ZHENG Yan,LI Jian,LI Zhi,ZHANG Jun-peng. Optimized Algorithm Based on Random Filter for Measurement Matrices. Journal of Chinese Computer Systems, 2016, 37(3): 632-636.