Underwater Non-stationary Acoustic Signal Detection Based on the STHOC Noise Suppression

Underwater non-stationary acoustic signal detection has a variety of applied research in marine, military and civilian fields. Moreover, the signal has strong randomness and short-term characteristics. Usually, the background noise sources of the actual signal detection environment have diversity and complexity. In the marine environment, signal detection methods like the short-time Fourier transform (STFT) may encounter uncertainties and limitations when dealing with unknown interferences such as Gaussian white noise and colored noise. In order to suppress marine background noise and obtain the best signal-to-noise ratio (SNR) for underwater acoustic observation signals detected in the frequency domain, this paper proposes a noise suppression detection and analysis method that combines short-term signal stabilization and higher-order cumulants. Through marine environment simulation and comparison experiments, the noise suppression effect of short-term higher-order cumulants (STHOCs) on Gaussian background noise under different SNR conditions and the performance improvement to signal frequency-domain (Fre-SNR) is analyzed. The processing results of experimental case data show that the frequency-domain noise energy standard deviation (Fre-STD) of the target signal is reduced. Compared to the STFT algorithm, the STHOC algorithm exhibits enhanced noise suppression and signal detection capabilities for target signals B and C. Finally, building upon the aforementioned analysis, the efficacy and superiority of the STHOC in detecting non-stationary underwater acoustic signals are further substantiated.