Assessing the relative detection performance of matched-field processing and conventional beamforming

Abstract

The primary advantage of matched-field processing (MFP) over conventional plane-wave beamforming is the ability to localize an acoustic source in a single data snapshot; for example, MFP can help to decide whether or not the source is submerged. It is to be expected that MFP will also provide improved detection performance over the plane-wave beamformer when the acoustic pressure field at the array exhibits the complicated behavior that is typically observed in the oceanic waveguide. Accurately quantifying the performance gain is difficult, owing to analytical problems in determining the exact performance of a detector that thresholds many correlated search cells. However, on the basis of simple statistical arguments concerning the effective dimensionality of the search spaces, it is claimed that a rough comparison of detector performance is provided by the ratio of the quadratic forms that are used as the decision variables. This ratio is computed for several array and ocean scenarios, including both a vertical line array in shallow water and a horizontal line array in deep water. Two-dimensional plots (depth and range) are used to illustrate the detection gain that can be provided by matched-field processing.