Electronics Optics & Control, Volume. 27, Issue 12, 45(2020)

An SVD Noise Reduction Method Using Fractal Dimension of Noise Signal

SHENG Pei, XU Aiqiang, CUI Weicheng, and JIANG Yi
Author Affiliations
  • [in Chinese]
  • show less

    To solve the problem that the determination of the order of an effective rank in Singular Value Decomposition (SVD) algorithm is not convincing enough and it is difficult to be applied in engineering, a new method utilizing noise characteristics is proposed.First, the singular value sequence is obtained by phase space reconstruction.Secondly, the fractal dimension of the filtered signal corresponding to each point is obtained.Finally, the optimal order of an effective rank is determined by the actual fractal dimension of noise.Simulation examples show the effectiveness of the proposed method, and it is compared with two typical algorithms to illustrate its advantages resulted from making full use of noise characteristics, and the variation of the algorithm performance with SNR is given.

    Tools

    Get Citation

    Copy Citation Text

    SHENG Pei, XU Aiqiang, CUI Weicheng, JIANG Yi. An SVD Noise Reduction Method Using Fractal Dimension of Noise Signal[J]. Electronics Optics & Control, 2020, 27(12): 45

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category:

    Received: Nov. 21, 2019

    Accepted: --

    Published Online: Jan. 14, 2021

    The Author Email:

    DOI:10.3969/j.issn.1671-637x.2020.12.010

    Topics