Deep Learning-Based Method for Key Signal Recognition during Underwater Explosions

ZHOU Zhenxian; HONG Feng; XU Weijie; ZHANG Tao; CHEN Feng

doi:10.11993/j.issn.2096-3920.2023-0146

Volume 32 Issue 4

Aug 2024

Turn off MathJax

Article Contents

Article Navigation > Journal of Unmanned Undersea Systems > 2024 > 32(4): 739-748

ZHOU Zhenxian, HONG Feng, XU Weijie, ZHANG Tao, CHEN Feng. Deep Learning-Based Method for Key Signal Recognition during Underwater Explosions[J]. Journal of Unmanned Undersea Systems, 2024, 32(4): 739-748. doi: 10.11993/j.issn.2096-3920.2023-0146

Citation:

ZHOU Zhenxian, HONG Feng, XU Weijie, ZHANG Tao, CHEN Feng. Deep Learning-Based Method for Key Signal Recognition during Underwater Explosions[J]. Journal of Unmanned Undersea Systems, 2024, 32(4): 739-748. doi: 10.11993/j.issn.2096-3920.2023-0146

Citation:

PDF( 1732 KB)

Deep Learning-Based Method for Key Signal Recognition during Underwater Explosions

doi: 10.11993/j.issn.2096-3920.2023-0146

1.
Shanghai Acoustics Laboratory, Chinese Academy of Sciences, Shanghai 201815, China
2.
University of Chinese Academy of Sciences, Beijing 100049, China

Received Date: 2023-11-10
Accepted Date: 2024-01-08
Rev Recd Date: 2023-12-27

Available Online: 2024-08-06

Abstract

Abstract

The amount of data collected from underwater explosion tests is enormous, which is mixed with a large amount of useless data. To protect the data from the effects of the explosion, it is crucial to prioritize the recognition and storage of key data during the test. In response to this, a key signal recognition model that combined feature extraction methods with deep learning models was proposed to improve the accuracy of key signal recognition. Firstly, different preprocessing methods for removing trend components from underwater explosion acceleration signals were studied. Existing test results demonstrated that wavelet packet decomposition, empirical mode decomposition, and high-pass filtering could significantly enhance the model’s recognition performance. Secondly, to make the extracted features more conducive to distinguishing between explosion and non-explosion segments, a feature extraction method based on the inter-class variance ratio for underwater explosion acceleration signals was proposed. According to the actual measured underwater explosion acceleration signal data, it was found that compared to Log Mel features, the proposed features improved classification accuracy by approximately 4.92% using the K-means method. Finally, the ECAPA-TDNN model incorporating the SE-Res2Block module was introduced, ensuring better recognition accuracy. With the proposed features as input, the recognition accuracy reached 99.31%.
- underwater explosion,
- feature extraction,
- deep learning,
- key signal recognition

FullText(HTML)

References(25)

References

[1]	刘建湖. 舰船非接触水下爆炸动力学的理论与应用[D]. 无锡: 中国船舶科学研究中心, 2002.
[2]	CHUNG J, SEO Y Y, YOUNG S. Shin dynamic and whipping response of the surface ship subjected to underwater explosion: experiment and simulation[J]. Ships and Offshore Structures, 2020, 15(10): 1129-1140. doi: 10.1080/17445302.2019.1706924
[3]	JIN P, XU H L, WANG H C, et al. On the seismic source function of an underwater explosion[J]. Geophysical Journal International, 2023, 232(1): 485-503.
[4]	CHUNG J, SEO Y, SHIN Y S. Dynamic and whipping response of the surface ship subjected to underwater explosion: experiment and simulation[J]. Ships and Offshore Structures, 2019, 15(10): 1129-1140.
[5]	杜志鹏, 张磊, 谌勇, 等. 泡沫覆盖层对水下爆炸气泡射流防护机理缩比试验研究[J]. 应用数学和力学, 2022, 43(5): 569-576. DU Z P, ZHANG L, CHEN Y, et al. Reduced-scale experiment study on the protective mechanism of foam coating against underwater explosion bubble jet[J]. Applied Mathematics and Mechanics, 2019, 43(5): 569-576.
[6]	范志强, 马宏昊, 沈兆武, 等. 水下连续脉冲冲击波的声学特性[J]. 爆炸与冲击, 2013, 33(5): 501-506. FAN Z Q, MA H H, SHEN Z W, et al. Acoustic characteristics of underwater continuous pulse shock waves[J]. Explosion and Shock Waves, 2013, 33(5): 501-506.
[7]	REZAEE M, TARAGHI O A. Improving empirical mode decomposition for vibration signal analysis[J]. Proceedings of the Institution of Mechanical Engineers, Part C. 2017, 231(12): 2223-2234.
[8]	蒲坚, 崔硕, 黄丹, 等. 山岭隧道爆破振动信号小波包及能量分析[J]. 交通科技, 2018(2): 61-65. PU J, CUI S, HUANG D, et al. Wavelet packet and energy analysis of vibration signals in mountain tunnel blasting[J]. Journal of Traffic Science and Technology, 2018(2): 61-65.
[9]	ROGÉRIO T, WENDERSON N L, PAULO R A, et al. Digital signal processing for self-vibration monitoring in grinding: A new approach based on the time-frequency analysis of vibration signals[J]. Measurement, 2019, 145: 71-83. doi: 10.1016/j.measurement.2019.05.079
[10]	李洪涛. 基于无线数据传输的水下爆炸压力遥测系统硬件设计与实现[D]. 长沙: 国防科学技术大学, 2002.
[11]	贾振华, 王文廉. 瞬态压力测试系统中信号识别触发的设计与实现[J]. 火工品, 2016(1): 57-60. JIA Z H, WANG W L. Design and implementation of signal recognition trigger in transient pressure testing system[J]. Chinese Journal of Explosives & Propellants, 2016(1): 57-60.
[12]	PRIOR M K. An optimization approach to the automatic identification of signals originating from underwater explosions[J]. The Journal of the Acoustical Society of America, 2008, 123(5): 3900.
[13]	ZHA X, PENG H, QIN X, et al. A deep learning framework for signal detection and modulation classification[J]. Sensors, 2019, 19: 4042.
[14]	ZHONG M, MANUEL C, RAHUL D, et al. Beluga whale acoustic signal classification using deep learning neural network models[J]. J. Acoust. Soc. Am, 2020, 147(3): 1834-1841. doi: 10.1121/10.0000921
[15]	JÚLIO d C V F, NATANAEL N d M J, JOSÉ M d S. Deep learning models for passive sonar signal classification of military data[J]. Remote Sens, 2022, 14: 2648. doi: 10.3390/rs14112648
[16]	王乾勋, 闫明, 杜志鹏, 等. 趋势项误差的低频极限特征理论模型与处理方法研究[J]. 振动与冲击, 2018, 37(12): 239-243. WANG Q X, YAN M, DU Z P, et al. Research on low-frequency limit characteristic theoretical model and processing method of trend item error[J]. Journal of Vibration and Shock, 2018, 37(12): 239-243.
[17]	徐卓飞, 刘凯. 基于极值符号序列分析的EMD端点效应处理方法[J]. 振动、测试与诊断, 2015, 35(2): 309-315, 400. XU Z F, Liu K. Endpoint effect handling method for EMD based on extremum sign sequence analysis[J]. Vibration, Testing and Diagnosis, 2015, 35(2): 309-315, 400.
[18]	董晨懿, 陈梦英, 许伟杰, 等. 一种改进的水下爆炸冲击波信号修正方法[J]. 声学技术, 2022, 41(3): 376-381. DONG C Y, CHEN M Y, XU W J, et al. An improved correction method for underwater explosion shock wave signals[J]. Acoustic Technology, 2022, 41(3): 376-381.
[19]	STEVENS S S, VOLKMANN J, NEWMAN E B. The mel scale equates the magnitude of perceived differences in pitch at different frequencies[J]. Journal of the Acoustical Society of America, 1937, 8(3): 185-190. doi: 10.1121/1.1915893
[20]	伊鑫, 曲爱华. 基于Welch算法的经典功率谱估计的Matlab分析[J]. 现代电子技术, 2010, 33(3): 7-9. YI X, QU A H. Matlab analysis of classical power spectral density estimation based on welch algorithm[J]. Modern Electronic Technology, 2010, 33(3): 7-9.
[21]	张志伟, 杨可林, 冯志常, 等. 基于小波包倒谱系数和ECAPA-TDNN的调度说话人确认研究[J]. 山东电力技术, 2023, 50(2): 52-57. ZHANG Z W, YANG K L, FENG Z C, et al. Research on speaker verification in power dispatching based on wavelet packet cepstral coefficient and ECAPA-TDNN[J]. Shandong Electric Power, 2023, 50(2): 52-57.
[22]	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition(CVPR). Las Vegas, NV, USA: IEEE, 2016: 770-778.
[23]	HU J, SHEN L, SUN G. Squeeze and excitation networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition(CVPR). Salt Lake City, UT, USA: IEEE, 2018: 7132-7141.
[24]	DESPLANQUES B, THIENPONDT J, DEMUYNCK K. ECAPA-TDNN: Emphasized channel attention, propagation and aggregation in TDNN based speaker verification[EB/OL]. (2020-05-15)[2023-10-10]. http://arxiv.org/abs/2005.07143v3.
[25]	吴夙慧, 成颖, 郑彦宁, 等. K-means算法研究综述[J]. 现代图书情报技术, 2011(5): 28-35. WU S H, CHENG Y, ZHENG Y N, et al. A review of K-means algorithm research[J]. Modern Library and Information Technology, 2011(5): 28-35.