Anti-noise-folding regularized subspace pursuit recovery algorithm for noisy sparse signals

Xianjun Yang, Qimei Cui, Eryk Dutkiewicz, Xiaojing Huang, Xiaofeng Tao, Gengfa Fang

    Research output: Chapter in Book/Report/Conference proceedingConference proceeding contributionpeer-review

    3 Citations (Scopus)

    Abstract

    Denoising recovery algorithms are very important for the development of compressed sensing (CS) theory and its applications. Considering the noise present in both the original sparse signal x and the compressive measurements y, we propose a novel denoising recovery algorithm, named Regularized Subspace Pursuit (RSP). Firstly, by introducing a data pre-processing operation, the proposed algorithm alleviates the noise-folding effect caused by the noise added to x. Then, the indices of the nonzero elements in x are identified by regularizing the chosen columns of the measurement matrix. Afterwards, the chosen indices are updated by retaining only the largest entries in the Minimum Mean Square Error (MMSE) estimated signal. Simulation results show that, compared with the traditional orthogonal matching pursuit (OMP) algorithm, the proposed RSP algorithm increases the successful recovery rate (and reduces the reconstruction error) by up to 50% and 86% (35% and 65%) in high noise level scenarios and inadequate measurements scenarios, respectively.

    Original languageEnglish
    Title of host publicationIEEE Wireless Communications and Networking Conference, WCNC
    Place of PublicationPiscataway, NJ
    PublisherInstitute of Electrical and Electronics Engineers (IEEE)
    Pages275-280
    Number of pages6
    ISBN (Electronic)9781479930838
    DOIs
    Publication statusPublished - 2014
    Event2014 IEEE Wireless Communications and Networking Conference, WCNC 2014 - Istanbul, Turkey
    Duration: 6 Apr 20149 Apr 2014

    Other

    Other2014 IEEE Wireless Communications and Networking Conference, WCNC 2014
    CountryTurkey
    CityIstanbul
    Period6/04/149/04/14

    Fingerprint Dive into the research topics of 'Anti-noise-folding regularized subspace pursuit recovery algorithm for noisy sparse signals'. Together they form a unique fingerprint.

    Cite this