A new recognition and classification algorithm of underwater acoustic signals based on multi-domain features combination

Abstract

A new algorithm is proposed in this paper to recognize and classify underwater acoustic signals including underwater acoustic communication signals, active sonar signals, ship-radiating noise and burst ocean ambient noise. The paper firstly analyzed the features of underwater acoustic signals in time domain, frequency domain and cyclostationarity domain. At the same time, the way of extracting features about these signals is researched. This algorithm combines the features above and chooses a decision tree classifier to realize the recognition and classification of different underwater acoustic signals. Finally, the effectiveness in different multipath channels is evaluated by simulation. The results have shown that this algorithm is capable of recognizing and classifying different underwater acoustic signals with better performance than algorithms using features in only one domain; and, its performance is robust in a multipath fading environment.