Statistical characterization of shallow-water bistatic sonar reverberation and noise

Abstract

Shallow-water, low-frequency bistatic sonar measurements often exhibit high noise caused by bathymetric reverberation and other sources. Understanding the noise statistics is an essential first step in developing filters that separate noise clutter from desired signals. This paper characterizes matched filter outputs from a bistatic receiver for a shallow-water environment for several waveforms. Data dominated by ambient noise are compared to reverberation from identifiable bathymetric rcturns. The direct blast reverberation tail is excluded. The measurements include various types of envelope time series power and powerspectral statistics. This analysis shows that ambient noise can be approximately explained using synthetic noisc. Bathymetric returns have different statistics and are not predictable by synthetic noise.