Sonar echo signal processing based on Convolution Blind source separation

Abstract

The existence of noise, sound scattering and reverberation in active sonar echo signal makes it very difficult to extract the target echo signal. This paper uses the convolution blind source separation algorithm to process the active sonar echo signal. The signal is transferred to the frequency-domain by sliding the STFT. The JADE blind source separation algorithm is used in the frequency domain, and the signal is finally inverted to the time domain through the ISTFT. The separation effect is compared and evaluated through similarity coefficients, time-domain waveforms, and time-frequency diagrams. The simulation results show that this method has a good effect on the separation of active sonar echo signals.