Page 102 - 《应用声学》2019年第6期

第 38 卷 第 6 期 Vol. 38, No. 6
2019 年 11 月 Journal of Applied Acoustics November, 2019

⋄ 研究报告 ⋄

基于深度神经网络的水声信号恢复方法研究 ∗

王全东 1,2 郭良浩 1† 闫 超 1

(1 中国科学院声学研究所 声场声信息国家重点实验室 北京 100190)
(2 中国科学院大学 北京 100049)

摘要 针对干扰或噪声环境下水声目标信号难以获取的问题，该文提出研究基于深度神经网络的自适应水声
被动信号波形恢复方法。在单阵元情况下，该方法提取对数功率谱特征作为输入，采用深度神经网络回归模型
自适应学习目标信号的自身特征，输出降噪后的对数功率谱特征并还原时域波形。在多阵元情况下，提出阵列
深度神经网络降噪方法，将部分或全部阵元特征拼接为长向量作为输入，从而利用空域信息。为全面利用阵列
丰富的时频域信息，该文提出一种两阶段特征融合深度神经网络，在第一阶段将阵列分为若干个子阵，将每个
子阵分别用阵列深度神经网络进行处理，在第二阶段将第一阶段的各子阵处理结果与阵列接收信号同时输入
一个深度神经网络进行融合学习。实验表明，所提出的单阵元和两阶段融合深度神经网络取得了显著优于常
规波束形成的恢复结果，能够准确估计目标信号波形和功率并显著提高输出信噪比。
关键词 水声被动降噪，深度神经网络，阵列处理，特征映射
中图法分类号: O427.9 文献标识码: A 文章编号: 1000-310X(2019)06-1004-11
DOI: 10.11684/j.issn.1000-310X.2019.06.014

Underwater acoustic target waveform recovery based on deep neural networks

WANG Quandong 1,2 GUO Lianghao 1 YAN Chao 1

(1 State Key Laboratory of Acoustics, Institute of Acoustics, Chinese Academy of Sciences, Beijing 100190, China)
(2 University of Chinese Academy of Sciences, Beijing 100049, China)

Abstract A deep neural network (DNN) based method is proposed to recover underwater acoustic target
signal waveform under noise. In single sensor condition, the log-power spectral (LPS) feature is extracted as
input, and a DNN regression model is employed to adaptively learn the inherent pattern of target signal and
output the enhanced LPS to recover the waveform. In multi-sensor condition, an array-based DNN which uses
the concatenated feature from partial or all sensors as input is proposed to exploit spatial information. To fully
use the rich temporal and spatial information from the array, we propose a two-stage DNN. In the first stage, the
array is split into sub-arrays and each sub-array is processed by an array-based DNN, while in the second stage,
the sub-array enhanced features and noisy array features are input to a DNN for integration. Experiments show
that our single-sensor and two-stage DNN achieved far better recovery results than conventional beamforming,
can accurately recover the target waveform and power and significantly improve the output signal-to-noise
ratio.
Key words Underwater acoustic signal recovery, Deep neural network, Array processing, Feature mapping

2019-01-28 收稿; 2019-03-25 定稿
国家自然科学基金项目 (61571436)
∗
作者简介: 王全东 (1991- ), 男, 黑龙江哈尔滨人, 博士研究生, 研究方向: 信号与信息处理。
† 通讯作者 E-mail: glh2002@mail.ioa.ac.cn