Proceedings of the 13th EAI International Conference on Mobile Multimedia Communications, Mobimedia 2020, 27-28 August 2020, Cyberspace

Research Article

Research on Non-stationary Blind Separation Method for Unfixed Signal Based on Extended Joint Diagonalization

Download344 downloads
  • @INPROCEEDINGS{10.4108/eai.27-8-2020.2294590,
        author={jun  li and Xin-Yuan  Chen},
        title={Research on Non-stationary Blind Separation Method for Unfixed Signal Based on Extended Joint Diagonalization },
        proceedings={Proceedings of the 13th EAI International Conference on Mobile Multimedia Communications, Mobimedia 2020, 27-28 August 2020, Cyberspace},
        publisher={EAI},
        proceedings_a={MOBIMEDIA},
        year={2020},
        month={11},
        keywords={extended joint diagonalization; nonstationary signal; underfixed blind separation; number of iterations},
        doi={10.4108/eai.27-8-2020.2294590}
    }
    
  • jun li
    Xin-Yuan Chen
    Year: 2020
    Research on Non-stationary Blind Separation Method for Unfixed Signal Based on Extended Joint Diagonalization
    MOBIMEDIA
    EAI
    DOI: 10.4108/eai.27-8-2020.2294590
jun li,*, Xin-Yuan Chen1
  • 1: College of Artificial Intelligence,Jiangxi Institute of Applied Science and Technology
*Contact email: lijun73812@163.com

Abstract

In order to solve the problem of signal instability caused by the number of iterations of the traditional nonstationary signal underfixed blind separation method, a nonstationary signal underfixed blind separation method based on extended combined diagonalization method is studied. Firstly, the second-order correlation matrix is extended to the fourth-order accumulation by the extended joint diagonalization method, the linear time-frequency variation of the nonstationary signal is analyzed, the uncertain part of the nonstationary signal is calculated, the function expression of the source signal is obtained, and the time-domain and frequency-domain changes in the function expression of the iterative source signal are obtained. The experimental results show that compared with the traditional separation method, the nonstationary signal underfixed blind separation method based on extended joint diagonalization method has the least number of iterations, and the isolated signal is the most stable.