Journal of Northeastern University Natural Science ›› 2019, Vol. 40 ›› Issue (7): 920-924.DOI: 10.12068/j.issn.1005-3026.2019.07.002

• Information & Control • Previous Articles     Next Articles

Underdetermined Blind Source Separation Algorithm Based on Directional Amplitude Ratio

JI Ce, JIANG Yu-tian   

  1. School of Computer Science & Engineering, Northeastern University, Shenyang 110169, China.
  • Received:2018-06-13 Revised:2018-06-13 Online:2019-07-15 Published:2019-07-16
  • Contact: JIANG Yu-tian
  • About author:-
  • Supported by:
    -

Abstract: In order to improve the estimation accuracy of mixed matrix in underdetermined blind source separation, an underdetermined blind source separation algorithm based on directional amplitude ratio of mixed signal data in frequency domain was proposed. In order to make full use of the phase information of data points of complex mixed signals, the directional amplitude ratio of complex signals was introduced, and the screening of single source points was accurately implemented by the variance, mean and distribution density of directional amplitude ratio of the complex mixed signals. The single source points was disposed with unitization and projection, the ratio of column elements of a mixed matrix was obtained through clustering analysis to estimate the mixed matrix. Finally, the matching pursuit algorithm was used to reconstruct the source signal. The simulation results show that the algorithm proposed can obtain more precise mixed matrix and separated signals compared with the comparison algorithms.

Key words: underdetermined blind source separation, directional amplitude ratio, single source points, clustering analysis, matching pursuit algorithm

CLC Number: