Characteristic Acquiring And Analyzing Of Underwater Transient Signal

Underwater transient signals, such as the sound produced by the torpedo or other objects entering the water from air at a certain speed, and target classification are all important for torpedo alert. These become the hot and difficult field because the underwater transient signals and underwater background noise are complex. In this paper, the characteristics of the sound produced by the torpedo or other objects entering water and classing these objects are important.In this paper, the process of objects entering water and the composition of the sound produced by the torpedo or other objects entering water are researched. Hilbert-Huang Transform, refer to as HHT, theory is used to deal with the non-stationary signals. HHT theory is researched and simulated in this paper, it is better than Wavelet Transform in dealing with underwater transient signals.Energy test method and high order statistics method are applied for detecting the impact pulse signals produced by the entrance of the object into water. A series experiments of objects enter water were completed in muffled tank and Songhua Lake. At the same time, the signals collected in these experiments are analyzed by HHT theory.Fuzzy Adaptive Resonance Theory Mapping, referred to as FARTMAP,neural network which is stable and adaptive at the same time, is simulated in thispaper. The signals collected in Songhua Lake are analyzed by FARTMAP in order to class the targets entering water from air at a certain speed.