Multi-channel back-propagation localization of an acoustic autonomous underwater vehicle and the effects of physical oceanographic variability

A multi-channel back-propagation approach was used to acoustically localize an autonomous underwater vehicle (AUV), which was out of range for bottom-lock navigation and had resorted to unreliable dead-reckoning. This opportunistic study was performed as part of the New England Shelf Break Acoustics experiment in spring 2021, when an AUV was deployed within a network of acoustic transceiver moorings and ship-towed sources. The AUV was equipped with a 2.5-4.5 kHz transducer and towed a 16-channel hydrophone array. This localization method involves ray back-propagation of direct and surface reflected arrivals originating from the AUV source and received on nearby mooring hydrophones. Additional back-propagation of ship-towed source signals received on the AUV array is used to determine vehicle azimuth. The effects of mooring tilt are analyzed and incorporated into the localization result to minimize error. Localization uncertainty related to source-receiver geometry and physical oceanographic variability is investigated, and a geometric model for characterizing uncertainty is presented. The final localization result is then compared to that of other localization methods, including AUV dead-reckoning and multi-channel back-propagation without a depth constraint.