Classification of underwater acoustic transients by artificial neural networks

Abstract

The goal of the research described was to study the feasibility of using artificial neural networks to recognize (or classify) acoustic transient signals that have been propagated through an ocean environment, including surface and bottom effects. The networks were tested on signals propagated to 25 different receiver sites by the time-domain parabolic equation model. Despite the interference effects from surface and bottom reflections/refractions, the classification accuracy was about 90% in the noise-free case. Classification in the presence of noise is reduced. However, the redundancy provided by the multiple receivers in most cases allows the network to correctly classify all signals from sources on which it was trained. It shows a robustness in the presence of unknown signals not shown by the nearest-neighbor classifier.<>