Deep transfer learning for underwater direction of arrival using one vector sensora)

被引:40
作者
Cao, Huaigang [1 ,3 ]
Wang, Wenbo [1 ,3 ]
Su, Lin [1 ,3 ]
Ni, Haiyan [1 ,3 ]
Gerstoft, Peter [2 ]
Ren, Qunyan [1 ,3 ]
Ma, Li [1 ,3 ]
机构
[1] Chinese Acad Sci, Inst Acoust, Key Lab Underwater Acoust Environm, Beijing 100190, Peoples R China
[2] Univ Calif San Diego, Scripps Inst Oceanog, NoiseLab, La Jolla, CA 92093 USA
[3] Univ Chinese Acad Sci, Beijing 100049, Peoples R China
基金
中国国家自然科学基金;
关键词
CONVOLUTIONAL NEURAL-NETWORK; SOURCE LOCALIZATION; CLASSIFICATION; AZIMUTH; OCEAN;
D O I
10.1121/10.0003645
中图分类号
O42 [声学];
学科分类号
070206 ; 082403 ;
摘要
A deep transfer learning (DTL) method is proposed for the direction of arrival (DOA) estimation using a single-vector sensor. The method involves training of a convolutional neural network (CNN) with synthetic data in source domain and then adapting the source domain to target domain with available at-sea data. The CNN is fed with the cross-spectrum of acoustical pressure and particle velocity during the training process to learn DOAs of a moving surface ship. For domain adaptation, first convolutional layers of the pre-trained CNN are copied to a target CNN, and the remaining layers of the target CNN are randomly initialized and trained on at-sea data. Numerical tests and real data results suggest that the DTL yields more reliable DOA estimates than a conventional CNN, especially with interfering sources.
引用
收藏
页码:1699 / 1711
页数:13
相关论文
共 40 条
[1]  
[Anonymous], 2015, ACS SYM SER
[2]  
[Anonymous], 1991, ILR
[3]   Machine learning in acoustics: Theory and applications [J].
Bianco, Michael J. ;
Gerstoft, Peter ;
Traer, James ;
Ozanich, Emma ;
Roch, Marie A. ;
Gannot, Sharon ;
Deledalle, Charles-Alban .
JOURNAL OF THE ACOUSTICAL SOCIETY OF AMERICA, 2019, 146 (05) :3590-3628
[4]   Sound source ranging using a feed-forward neural network trained with fitting-based early stopping [J].
Chi, Jing ;
Li, Xiaolei ;
Wang, Haozhong ;
Gao, Dazhi ;
Gerstoft, Peter .
JOURNAL OF THE ACOUSTICAL SOCIETY OF AMERICA, 2019, 146 (03) :EL258-EL264
[5]   Convolutional neural network for single-sensor acoustic localization of a transiting broadband source in very shallow water [J].
Ferguson, Eric L. ;
Williams, Stefan B. ;
Jin, Craig T. .
JOURNAL OF THE ACOUSTICAL SOCIETY OF AMERICA, 2019, 146 (06) :4687-4698
[6]  
Ferguson EL, 2017, INT CONF ACOUST SPEE, P2657, DOI 10.1109/ICASSP.2017.7952638
[7]   Identifiability in array processing models with vector-sensor applications [J].
Hochwald, B ;
Nehorai, A .
IEEE TRANSACTIONS ON SIGNAL PROCESSING, 1996, 44 (01) :83-95
[8]   Source localization using deep neural networks in a shallow water environment [J].
Huang, Zhaoqiong ;
Xu, Ji ;
Gong, Zaixiao ;
Wang, Haibin ;
Yan, Yonghong .
JOURNAL OF THE ACOUSTICAL SOCIETY OF AMERICA, 2018, 143 (05) :2922-2932
[9]  
Jacobsen MH, 2013, CLASS CONT SOC THEOR, P13
[10]  
Jensen FB, 2011, MOD ACOUST SIGN PROC, P337, DOI 10.1007/978-1-4419-8678-8_5