Spectral estimation of cavitation related narrow-band ship radiated noise based on fractional lower order statistics and multiple signal classification

Narrow-band spectral components of ship radiated noise are of crucial importance for sonar operators in order to detect and classify possible targets. Classical spectral estimation techniques have been widely used to fulfill the requirements for narrow-band signal detection. However, classical methods may not be sufficient when signal-to-noise ratio is relatively low. In this paper we propose to utilize the subspace based multiple signal classification for the cavitation related narrow-band ship radiated noise detection. In addition we propose the use of fractional lower order statistics instead of second order statistics when the ship radiated noise exhibits an impulsive behavior leading to drifting apart from Gaussianity. We present the results of a real ship noise data gathered from the Strait of Istanbul. We also investigate and present the 1/f-type noise characteristics of a sample ship noise segment and discuss how we can use this information for robust estimation of cavitation related narrow-band ship noise spectrum.