Coronavirus Information for the UC San Diego Community

Our leaders are working closely with federal and state officials to ensure your ongoing safety at the university. Stay up to date with the latest developments. Learn more.

Joint towed array shape and direction of arrivals estimation using sparse Bayesian learning during maneuvering

TitleJoint towed array shape and direction of arrivals estimation using sparse Bayesian learning during maneuvering
Publication TypeJournal Article
Year of Publication2020
AuthorsZheng Z., Yang T.C, Gerstoft P, Pan X.
Volume147
Pagination1738-1751
Date Published2020/03
Type of ArticleArticle
ISBN Number0001-4966
Accession NumberWOS:000522105300001
Keywordsacoustics; Audiology & Speech-Language Pathology; framework; localization
Abstract

Large aperture towed arrays are widely used underwater to detect weak targets. During maneuvering, the beamformer performance degrades significantly if a wrong array configuration is assumed. Currently, engineering sensors and/or (augmented) acoustic sources are used to estimate the array element positions. The results are often inadequate depending on the number of measurements available. In this paper, an adaptive bow (AB) sparse Bayesian learning (SBL) algorithm is proposed, called ABSBL. Assuming the towed array follows a parabola shape during slow turns and treating the array bow as a hyperparameter in SBL, the bow and directions of arrival (DOAs) of the signals can be jointly estimated from the received acoustic data. Simulations show that ABSBL yields accurate estimates of the bow and target DOAs if the turning direction is known. ABSBL is applied to the MAPEX2000 data. The estimated array bow and DOA agrees with that estimated from relative time delays measured from acoustic pings and SBL, better than that estimated from the GPS data using the water-pulley model. The method can potentially be applied without engineering sensors.

DOI10.1121/10.0000920
Student Publication: 
No
Research Topics: 
sharknado