MULTICHANNEL SYSTEM FOR DETECTING HYDROACOUSTIC SIGNALS BASED ON INFORMATION FEATURES Russian patent published in 2024 - IPC G01S15/04 G01S7/52 

FIELD: hydroacoustics.

SUBSTANCE: invention relates to hydroacoustics, namely to devices for detecting hydroacoustic signals, and is aimed at detecting signals from a surface or underwater source with a low signal-to-noise ratio in conditions when interference is present in the form of random noise and/or is close to white noise. In a multichannel system for detecting hydroacoustic signals based on information features, such as entropy and statistical complexity, each channel comprises a memory unit associated with a fast Fourier transform unit (FFT) for storing averaged spectra obtained by averaging a given number of instantaneous spectra, output of which is connected to the averaged spectra processing unit to form normalized spectral values, connected to the input of the energy spectrum signal values accumulation and storage unit, equipped with a stack memory, with which there is a device for displaying processing results, which is configured to display a panoramic view of the target and discrete components detected in the signal, as well as to form an information image of the target. In the averaged spectra processing unit, when generating normalized spectral values in each window, a short-term energy spectrum is determined, for which the spectral power density of the signal is calculated depending on the number of spectral components in the FFT and the spectral decomposition of the signal in the frequency domain, and a discrete distribution of the spectral power density depending on the spectral energy for the spectral component with frequency ƒ_i, number of spectral components of FFT, after which for each obtained short-term spectrum of all windows, spectral entropy, imbalance and statistical complexity are calculated. To determine imbalance, different distance functions are selected.

EFFECT: high efficiency and reliability of detecting hydroacoustic signals with a low signal-to-noise ratio, as well as resistance to external noise based on spectral information characteristics such as entropy and statistical complexity.

1 cl, 13 dwg