Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Future Full-Ocean Deep SSPs Prediction based on Hierarchical Long Short-Term Memory Neural Networks

Jiajun Lu, Hao Zhang, Pengfei Wu, Sijia Li, Wei Huang

TL;DR

This work tackles the challenge of predicting future underwater sound speed profiles (SSPs) without on-site measurements. It introduces a hierarchical long short-term memory (H-LSTM) framework that treats SSPs at different depths as separate time series and trains depth-specific models, leveraging a rolling window to maximize historical information. Compared with mean-value, polynomial fitting, and BP neural networks, the H-LSTM achieves superior accuracy, with RMSEs as low as $0.5565$ m/s for full-ocean deep SSP predictions and consistently under $1$ m/s for monthly-average SSP across depths. The approach significantly enhances time efficiency for underwater PNT applications and enables reliable predictive capability for SSP dynamics over the full-depth ocean.

Abstract

The spatial-temporal distribution of underwater sound velocity affects the propagation mode of underwater acoustic signals. Therefore, rapid estimation and prediction of underwater sound velocity distribution is crucial for providing underwater positioning, navigation and timing (PNT) services. Currently, sound speed profile (SSP) inversion methods have a faster time response rate compared to direct measurement methods, however, most SSP inversion methods focus on constructing spatial dimensional sound velocity fields and are highly dependent on sonar observation data, thus high requirements have been placed on observation data sources. To explore the distribution pattern of sound velocity in the time dimension and achieve future SSP prediction without sonar observation data, we propose a hierarchical long short-term memory (H-LSTM) neural network for SSP prediction. By our SSP prediction method, the sound speed distribution could be estimated without any on-site data measurement process, so that the time efficiency could be greatly improved. Through comparing with other state-of-the-art methods, H-LSTM has better accuracy performance on prediction of monthly average sound velocity distribution, which is less than 1 m/s in different depth layers.

Future Full-Ocean Deep SSPs Prediction based on Hierarchical Long Short-Term Memory Neural Networks

TL;DR

This work tackles the challenge of predicting future underwater sound speed profiles (SSPs) without on-site measurements. It introduces a hierarchical long short-term memory (H-LSTM) framework that treats SSPs at different depths as separate time series and trains depth-specific models, leveraging a rolling window to maximize historical information. Compared with mean-value, polynomial fitting, and BP neural networks, the H-LSTM achieves superior accuracy, with RMSEs as low as m/s for full-ocean deep SSP predictions and consistently under m/s for monthly-average SSP across depths. The approach significantly enhances time efficiency for underwater PNT applications and enables reliable predictive capability for SSP dynamics over the full-depth ocean.

Abstract

The spatial-temporal distribution of underwater sound velocity affects the propagation mode of underwater acoustic signals. Therefore, rapid estimation and prediction of underwater sound velocity distribution is crucial for providing underwater positioning, navigation and timing (PNT) services. Currently, sound speed profile (SSP) inversion methods have a faster time response rate compared to direct measurement methods, however, most SSP inversion methods focus on constructing spatial dimensional sound velocity fields and are highly dependent on sonar observation data, thus high requirements have been placed on observation data sources. To explore the distribution pattern of sound velocity in the time dimension and achieve future SSP prediction without sonar observation data, we propose a hierarchical long short-term memory (H-LSTM) neural network for SSP prediction. By our SSP prediction method, the sound speed distribution could be estimated without any on-site data measurement process, so that the time efficiency could be greatly improved. Through comparing with other state-of-the-art methods, H-LSTM has better accuracy performance on prediction of monthly average sound velocity distribution, which is less than 1 m/s in different depth layers.
Paper Structure (35 sections, 19 equations, 10 figures, 5 tables)

This paper contains 35 sections, 19 equations, 10 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Structure of LSTM
  4. LSTM
  5. 1) Calculating output $f_t$ of forgetting gate:
  6. 2) Calculating output $i_t$ of input gate and candidate cell state $\tilde{C}_t$:
  7. 3) Updating cell state $C_t$:
  8. 4) Predicting $o_t$ of output gate and output $\hat{S}_{t+1}$ of hidden layer:
  9. H-LSTM
  10. Flowchart of H-LSTM for SSP Prediction
  11. Dataset pre-processing
  12. 1) Data layering by depth
  13. 2) Data segmentation
  14. 3) Further processing of hierarchical training sets
  15. H-LSTM neural network building
  16. ...and 20 more sections

Figures (10)

  • Figure 1: Unit structure of LSTM.
  • Figure 2: Folding-formed unit structure of LSTM.
  • Figure 3: The structure of H-LSTM.
  • Figure 4: Flow chart of H-LSTM for SSP prediction.
  • Figure 5: Spatial position of SSP samples.
  • ...and 5 more figures