Deep transfer learning for underwater direction of arrival using one vector sensor

Abstract

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.

Research on DCE-MRI Images Based on Deep Transfer Learning in Breast Cancer Adjuvant Curative Effect Prediction

“…Therefore, pathological examinations often occur before chemotherapy. After the operation and the later period of treatment, there is a lag effect to some extent [ 5 , 6 ]. Although there are many noninvasive examinations in the audience, dynamic contrast-enhanced MRI (DCE-MRI) has the most potential for development.…”

A multi‐step method for passive broadband source localisation using a single vector sensor

“…Sun et al [14,15] discussed the propagation characteristics of the vector field under the direct sound path in the deep ocean and proposed a method of nearsurface source localisation. Cao et al [16] applied a convolutional neural network to vector data and proposed a method of DOA estimation using an SVS based on deep transfer learning. Qi et al [17] used the interference cycle in the frequency domain to extract the source depth from signals recorded by an ocean-bottom seismometer (OBS) at a low frequency from 20 to 100 Hz.…”

“…Cao et al. [16] applied a convolutional neural network to vector data and proposed a method of DOA estimation using an SVS based on deep transfer learning. Qi et al.…”

Vector Sensor Steering-Dependent Performance in an Underwater Acoustic Communication Field Experiment

“…It is notorious and expected that most of the literature on vector sensors addresses the direction-finding issues since a compact collocated device can improve the gain of an ordinary pressure sensor or enhance the gain of a pressure array [ 7 , 8 , 9 , 10 ]. Much work has been carried out regarding direction of arrival (DoA) estimation using vector sensors, either by additive or multiplicative channel combining methods [ 7 , 11 , 12 , 13 ], exploiting signal and noise subspace domains [ 14 ], higher-order array manifold [ 15 ], or artificial intelligence [ 16 ]. Although researchers have deeply investigated DoA methods, the use of this information for UWAC is still little explored.…”