Geoacoustic Inversion in Shallow Water.

摘要

The ability to predict the sound field in shallow water is constrained by the knowledge of the geoacoustic properties of the bottom. The long term objectives of this research project are related to the investigation of experimental methods and inversion techniques for estimating parameters of geoacoustic models of the ocean bottom and the associated uncertainties in the model parameter values. The specific goals are to evaluate the performance of geoacoustic inversion techniques that have been developed for use in range- dependent shallow water environments, and synthesize the results obtained for characterizing the seabed from the SW06 and other recent experiments. The wider context of this research is to achieve improved sonar system performance through greater understanding of the physics of the interaction of sound with the ocean bottom. There are two objectives in this research report. The first is to compare the performance of geoacoustic inversion methods that have been developed for estimating geoacoustic models from acoustic field data. The inversion performance has been assessed previously in ONR Benchmarking workshops (Tolstoy et al., 1998; Chapman et al., 2003) that used simulated acoustic fields for several candidate shallow water waveguide environments. The hypothesis in this work is that data from the ONR SW06 experiment can serve as an experimental benchmark for assessing inversion performance against real data. This report compares the performance of geoacoustic inversion methods that were used to estimate parameters of geoacoustic profiles in SW06. These include matched field inversion, reflection coefficient and bottom loss inversion, and wavenumber extraction inversion. A critical issue in the many of the inversion methods is the impact of uncertainty in the knowledge of the ocean environment and the experimental geometry on inversion performance. A second objective in this work is to investigate approaches for estimating geoacoustic model parameters that are robust to environmental and experimental uncertainty.