Predictive modeling of future full-ocean depth SSPs utilizing hierarchical long short-term memory neural networks
| dc.contributor.author | Lu, Jiajun | |
| dc.contributor.author | Zhang, Hao | |
| dc.contributor.author | Wu, Pengfei | |
| dc.contributor.author | Li, Sijia | |
| dc.contributor.author | Huang, Wei | |
| dc.date.accessioned | 2024-06-07T22:53:30Z | |
| dc.date.available | 2024-06-07T22:53:30Z | |
| dc.date.issued | 2024 | |
| dc.description.abstract | The spatial-temporal distribution of underwater sound speed plays a critical role in determining the propagation mode of underwater acoustic signals. Therefore, rapid estimation and prediction of sound speed distribution are imperative for facilitating underwater positioning, navigation, and timing (PNT) services. While sound speed profile (SSP) inversion methods offer quicker response times compared to direct measurement methods, these methods often focus on constructing spatial sound velocity fields and heavily rely on sonar observation data, thus imposing stringent requirements on data sources. To delve into the temporal distribution pattern of sound speed and achieve SSP prediction without relying on sonar observation data, we introduce the hierarchical long short-term memory (H-LSTM) neural network for SSP prediction. Our method enables the estimation of sound speed distribution without the need for on-site data measurement, significantly enhancing time efficiency. Compared to other state-of-the-art approaches, the H-LSTM model achieves a root mean square error (RMSE) of less than 1 m/s in predicting monthly average sound speed distribution. Its prediction accuracy has improved several-fold over alternative methods, which validates the robust capability of our proposed model in predicting SSP. | |
| dc.description.reviewstatus | Reviewed | |
| dc.description.scholarlevel | Faculty | |
| dc.description.sponsorship | This research was funded by Natural Science Foundation of Shandong Province (ZR2023QF128), Laoshan Laboratory (LSKJ202205104), China Postdoctoral Science Foundation (2022M722990), Qingdao Postdoctoral Science Foundation (QDBSH20220202061), National Natural Science Foundation of China (NSFC:62271459), National Defense Science and Technology Innovation Special Zone Project: Marine Science and Technology Collaborative Innovation Center (22-05-CXZX-04-01-02), and the Fundamental Research Funds for the Central Universities, Ocean University of China (202313036). | |
| dc.identifier.citation | Lu, J., Zhang, H., Wu, P., Li, S., & Huang, W. (2024). Predictive modeling of future full-ocean depth SSPs utilizing hierarchical long short-term memory neural networks. Journal of Marine Science and Engineering, 12(6), 943. https://doi.org/10.3390/jmse12060943 | |
| dc.identifier.uri | https://doi.org/10.3390/jmse12060943 | |
| dc.identifier.uri | https://hdl.handle.net/1828/16597 | |
| dc.language.iso | en | |
| dc.publisher | Journal of Marine Science and Engineering | |
| dc.rights | Attribution 4.0 International | en |
| dc.rights.uri | http://creativecommons.org/licenses/by/4.0/ | |
| dc.subject | time series forecast | |
| dc.subject | predictive modeling | |
| dc.subject | full-ocean depth sound speed profile (SSP) | |
| dc.subject | hierarchical long short-term memory (H-LSTM) | |
| dc.subject.department | Department of Electrical and Computer Engineering | |
| dc.title | Predictive modeling of future full-ocean depth SSPs utilizing hierarchical long short-term memory neural networks | |
| dc.type | Article |