A fuzzy logic based underwater acoustic transient classifier

This paper discusses a fuzzy logic based system developed to detect and classify a series of real life underwater acoustic transients. Signals received in the ocean are highly contaminated by various sources. While a trained sonar operator may classify transients without too difficulty, automatic classification by traditional template matching does not give satisfactory results. Our classifier is based on the fuzzy logic theory which can incorporate human knowledge into the process of inference. In this way, we are able to consult experts and create a system which partially mimics human reasoning. The power of the classifier will be demonstrated with two classes of real life underwater acoustic transients.