Detection of phase‐coupled noises in the ocean

Abstract

Estimates of power spectral density are being used in a variety of signal processing applications. For the precise characterization of certain noise sources in the ocean, the phase blind power spectral estimation does not yield the required result. Estimates of the bispectrum and bicoherence have been found useful in detecting non‐Gaussianity and nonlinearity, in system identification, as well as transient signals. The power spectrum, in general, fails to quantify any nonlinear interactions between the component frequencies. Such interactions induced by second‐order nonlinearities give rise to certain phase relations called quadratic phase coupling (QPC). Bispectral analysis, which is based on the third‐order cumulant sequence of the signal, can reveal the presence of as well as provide a measure to quantify such phase couplings. Bispectral analyses of various noise data records collected from different ocean noise sources have been carried out and it is seen that the bispectrum is not identically zero, implying the non Gaussian nature of the noise data. This paper presents the bispectral estimation approach for identifying the nonlinearities in the underwater noise‐generating mechanisms as well as detecting the presence of quadratic phase couplings, thus leading to the identification of the noise sources.