Matched-field processing of 200-Hz cw data

Abstract

The 200-Hz cw data were collected in the NE Pacific with the Marine Physical Laboratory's acoustic vertical line array (a 900-m aperture with 120 sensors) and processed with matched-field processing (MFP) techniques. The environmental information collected during the experiment was used with a normal mode model assuming a range-independent medium to generate full wave field replica vectors for various array tilt angles. The MFP ambiguity surfaces were calculated using the Bartlett and minimum variance processors for two input signal-to-noise ratios (− 9.2 and 11.8 dB) and were characterized by localization and processing gain performance indices. The acoustic modeling leads to the detection and localization of the source near its expected location (20-m depth and 165-km range). While the minimum variance processor lacks robustness, the MFP Bartlett processor produces more processing gain than the conventional FFT beamformer and, for a 0.4° array tilt angle, has an estimated array gain only a few dB below its theoretical value of 21 dB. [Work supported by ONR.]