Range-Dependent Passive Source Localization Using Data from the Barents Sea Tomography Experiment.

• Main Author: Pierce, David D.
• Other Authors: NAVAL POSTGRADUATE SCHOOL MONTEREY CA
• Format: Text
• Language: English
• Published: 1996
• Subjects: Acoustic Detection and Detectors; Physical and Dynamic Oceanography; *UNDERWATER ACOUSTICS; *BARENTS SEA; *PASSIVE SONAR; MATHEMATICAL MODELS; ALGORITHMS; SIGNAL PROCESSING; UNCERTAINTY; SIGNAL TO NOISE RATIO; ANGLE OF ARRIVAL; DIRECTION FINDING; THESES; HIGH RESOLUTION; MATHEMATICAL FILTERS; WAVE PROPAGATION; SOUND TRANSMISSION; ACOUSTIC MEASUREMENT; NOISE REDUCTION; ACOUSTIC FILTERS; SOUND WAVES; ORDER STATISTICS; OCEAN ENVIRONMENTS; SONAR RECEIVERS; ACOUSTIC FIELDS; UNDERWATER SOUND SIGNALS; SOUND RANGING; DEPTH FINDING; MATCHED FIELD PROCESSING; Barents Sea
• Online Access: http://www.dtic.mil/docs/citations/ADA313862; http://oai.dtic.mil/oai/oai?&verb=getRecord&metadataPrefix=html&identifier=ADA313862

Matched-Field Processing (MFP) and Matched-Mode Processing (MMP) are two popular techniques for passively localizing an underwater acoustic emitter in range and depth. One major drawback of these techniques has been their sensitivity to uncertainty concerning the acoustic environment. Several methods for addressing this phenomenon have been proposed in the literature, with varying degrees of success. Achieving high-quality location estimates remains a problem except in simple range-independent experiments or numerical simulations. In this study, we demonstrate an approach for robust, accurate emitter localization in a highly range-dependent real environment using MMP. The main factors contributing to successful localization are: 1) use of the high-resolution Multiple Signal Classification (MUSIC) algorithm, which performs well even when only a few robust modes can be obtained by mode filtering; and 2) use of an acoustic propagation model incorporating mode coupling, which is able to generate accurate replica fields in a strongly range-dependent environment. A secondary objective of the study was to demonstrate the application of higher-order statistical estimation techniques to reduce noise effects. Our results indicate that these techniques show unacceptable sensitivity to noise- and model-induced estimation errors and require further refinement before they will be useful in the underwater acoustic localization problem.