Narrow-Band Short-Time Frequency-Domain Blind Signal Separation of Passive Sonar Signals

  • Authors:

  • Natanael N. Moura;Eduardo F. Simas Filho;José M. Seixas

  • Affiliations:

  • Signal Processing Laboratory, COPPE/Poli/UFRJ, Rio de Janeiro, Brazil;Signal Processing Laboratory, COPPE/Poli/UFRJ, Rio de Janeiro, Brazil and Federal Center for Technological Education, Bahia, Brazil;Signal Processing Laboratory, COPPE/Poli/UFRJ, Rio de Janeiro, Brazil

  • Venue:
  • ICA '09 Proceedings of the 8th International Conference on Independent Component Analysis and Signal Separation
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Sonar systems are very important for several military and civil navy applications. Passive sonar signals are susceptible to cross-interference from underwater acoustic sources (targets) present at different directions. In this work, a frequency-domain blind source separation procedure is proposed aiming at reducing cross-interferences, which may arise from adjacent signal source directions. As a consequence, target detection and classification may both be performed on cleaner data and one can expect an overall sonar efficiency improvement. As the underwater acoustic environment is time-varying, time-frequency transformation is performed using short-time windows. Original free of interference sources are estimated using ICA algorithms over narrow-band frequency-domain signals. It is shown that the proposed passive sonar signal processing approach attenuates in more than 10 dB the interference signals measured from two nearby directions and reduces the common background noise level in 7 dB.