Enlarging neural class detection capacity in passive sonar systems

Abstract

A neural discriminating analysis is used for classifying passive sonar signals. Preprocessed information from the amplitude spectra of the noise radiated from ships is projected onto only a few principal discriminating components for feeding the input nodes of the neural classifier. Envisaging practical applications, in which new incoming classes not known by the time of the training phase have to be detected in the production phase, a method is provided using the identification of outliers to trigger the arriving of a new class. Using experimental data, it is shown that up to 85% of patterns from an untrained class can be identified, without significant decrease on the efficiency of the classifer for classes known beforehand.

Análise de Amostras Sintéticas de Sinais de Sonar Passivo Geradas por Redes Neurais Generativas Adversariais

“…Depois, a representação de frequência do sinal é obtida aplicandose a transformada de Fourier (STFT -do inglês, Short Time Fourier Transform). Na sequência, o algoritmo TPSW (do inglês, Two Pass Split Window) [9] é utilizado para atenuar o ruído de fundo e normalizar o sinal. O resultado dessa análise pode ser exibido em um lofargrama, como visto na Fig.…”

Spectral analysis techniques for acoustic fingerprints recognition

“…In this way in [7] the information from sonar is preprocessed to extract the spectrum. In order to separate the kind of ship, PCA is used.…”

Acoustic Fingerprint Recognition Using Artificial Neural Networks

“…The acquired data were processed using Minimum Variance no Distortion Response (MVDR) techniques and the Multiple Signal Classification (MUSIC). Likewise, in [7] information from sonars is processed such that the amplitude spectrum is removed, then PCA is aimed to separate sets of boats, and the most significant components are used as input to a neural network applied for non-linear classification.…”