Underwater Acoustic Signal and Noise Modeling

TLDR: In this chapter, the four dimensions in which underwater acoustic signals can be categorized are introduced: time, frequency, consistency from observation to observation, and knowledge of structure.

Abstract: In this chapter, the four dimensions in which underwater acoustic signals can be categorized are introduced: time, frequency, consistency from observation to observation, and knowledge of structure. Recalling the remote-sensing application, the impact of propagation through an underwater acoustic channel on source-signal characterization is described in terms of its effect on signal amplitude and phase. Various representations of bandpass signals are presented, including the analytic signal, complex envelope, envelope and instantaneous intensity. Statistical models for sampled time-series data are obtained for signals and noise to support derivation and analysis of detection and estimation algorithms. Reverberation in active systems is characterized as a random process in order to describe its autocorrelation function and power spectral density. The effect on reverberation arising from the motion of the sonar platform or reverberation-source scatterers, known as Doppler spreading, is introduced and approximated. In addition to the standard Gaussian noise model, a number of heavy-tailed distributions are described including the K distribution, Poisson-Rayleigh, and mixture distributions. Standard statistical models for signals and signals-plus-noise are presented along with techniques for evaluating or approximating the probability of detection and probability of false alarm.