Reverberation suppression

Abstract

We propose a processing methodology for shallow-water reverberation

suppression in which acoustic propagation and scattering physics are used as a guide to develop simple, but still accurate representations for the received

data which are motivated by the ease of development of efficient signal-processing algorithms. The methodology is demonstrated by showing that temporally spread reverberation components can be accurately modeled using a reduced-rank representation and expansion by discrete prolate spheroidal sequences. These represenations lead to the application of the Principal Component Inverse (PCI) method and the reduced-rank generalized likelihood-ratio test (RR-GLRT) for suppressing reverberation and the development of an iterative scheme for estimating arrival times in the presence of temporally spread reverberation components. We also present a performance analysis of the RR-GLRT. Using both simulated and real data, it is shown that these methods perform well.