Abstract:

Classification of an echo at tactical ranges in a shallow-water environment is made difficult by dynamic shallow-water multipath. Stable deconvolution (or other) techniques can adapt the processing to the environment if reasonable and dynamic estimates of the environmental impulse response can be obtained. Technology that obtains real-time environmental impulse responses directly from the data record being analyzed acoustically is called ``through the sensor environmental estimation.'' This work develops a new technique to be added to the already existing arsenal of ``through the sensor'' techniques and is meant to complement the already-existing technology. This technique uses the ambient noise field and the fact that spatial diversity of the noise sources results in uncorrelated modes. As a consequence of this, the depth modal eigenfunctions for the waveguide can be obtained directly from an eigenvector decomposition of the data spatial covariance matrix, and the wave numbers can be obtained from beamforming these eigenfunctions. The estimated environment is tested here with passive matched-field simulations with limited success and with active matched-filter and biosonar simulations. [This work was supported by the Office of Naval Research. Technical management was provided by the Naval Research Laboratory, Stennis Space Center, Mississippi.]