Simulations and efficient prediction of matched-field processing degradation from mismatch in shallow water

Abstract

Mismatch between the replica and the measured fields in matched-field processing can result in a significant loss in detection and localization performance. Such mismatch effects may be estimated through computer intensive simulations. When this was done for vertical line arrays in Pacific and Arctic shallow water environments with typical uncertainties from conventional bathymetry and sediment sound-speed measurements, substantial mismatch effects were found. It is also possible to predict, with minimal calculation, the nature and, in some cases, the relative magnitude of mismatch effects. Simulated mismatch effects for both horizontal and vertical line array configurations in different shallow water environments agreed well in magnitude and nature with the predictions for environmental or geometric mismatch. Environmental mismatch studied included bathymetry and sound-speed errors while geometric mismatch included Doppler effects and array tilt. A consistent relationship was found between the degradation of the match at the true source location and the predictor. Different relationships were found for the best match on the ambiguity surface and the match at the true source location. The predictor was then employed to show how mismatch increases with frequency.