Estimators for Model-Based Passive Localization

Abstract : A comparative study is made of the performance of four different estimators as used in the matched-field technique of passive localization. The study is based on both real and synthesized data. In the synthesized data case, a comparison is made of the performance of the estimators for various signal-to- noise ratios. The four estimators studied are Bucker's Estimators, which can be thought of as a spatial matched filter, and three likelihood-type estimators. The results indicate that the matched-field type estimator has a slightly better signal-to-noise performance than the others, but rather poor sidelobe behaviour, whereas for the likelihood-type estimator the sidelobe behaviour is quite good. Keywords: Inverse problems; Matched-field processing; Model-based signal processing; Passive localization; Passive ranging; Hydrophones; Underwater acoustics; Mediterranean Sea.