Shijie Xu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Shijie Xu has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Guangming Li
National Innovation Institute of Defense Technology, Academy of Military Sciences, Beijing 100071, China
9 shared publications - Haocai Huang
Zhejiang University
9 shared publications - Pan Xu
College of Meteorology and Oceanography, National University of Defense Technology, Changsha 410073, China
5 shared publications - Fengyuan Yu
Ocean College, Zhejiang University, Zhoushan 316021, China
3 shared publications Rendong Feng
Ocean College, Zhejiang University, Zhoushan, China
3 shared publications- Zhengliang Hu
College of Meteorology and Oceanography, National University of Defense Technology, Changsha, China
3 shared publications
Latest Publications
- 29 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 4206513
Physics-Constrained Deep Neural Networks for Short-Term 3-D Flow Field Prediction Using Coastal Acoustic Tomography
Accurate prediction of three-dimensional (3-D) ocean flow fields is essential for understanding small-scale dynamic processes, improving environmental monitoring, and enhancing the performance of underwater observation and operation sys-tems. However, predictive modeling using coastal acoustic to-mography (CAT) data remains relatively unexplored. This study proposes a novel physics-constrained deep learning framework, termed PC-LSTM-MLP, to achieve the short-term prediction of 3-D ocean flow fields from CAT observations. The framework consists of three integrated components: a long short-term memory (LSTM) network for capturing temporal dependencies in range-averaged currents (RACs), a multilayer perceptron (MLP) for nonlinear mapping to spatial Fourier coefficients, and a physics-constrained (PC) module that imposes a horizontal incompressibility constraint to enhance the accuracy of the predicted 3-D flow fields. Experimental results on real CAT data demonstrate that the proposed PC-LSTM-MLP framework effec-tively captures complex spatiotemporal dynamics and provides accurate predictions up to 25 minutes ahead. The model achieves an average mean absolute error (MAE) of 0.0241 m/s and an average root mean square error (RMSE) of 0.0269 m/s in 3-D flow field prediction. This research shifts CAT analysis from traditional inversion methods to intelligent, prediction-oriented modeling, underscoring the significant potential of deep learning in ocean flow field prediction.
- 29 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 4206513
Physics-Constrained Deep Neural Networks for Short-Term 3-D Flow Field Prediction Using Coastal Acoustic Tomography
Accurate prediction of three-dimensional (3-D) ocean flow fields is essential for understanding small-scale dynamic processes, improving environmental monitoring, and enhancing the performance of underwater observation and operation sys-tems. However, predictive modeling using coastal acoustic to-mography (CAT) data remains relatively unexplored. This study proposes a novel physics-constrained deep learning framework, termed PC-LSTM-MLP, to achieve the short-term prediction of 3-D ocean flow fields from CAT observations. The framework consists of three integrated components: a long short-term memory (LSTM) network for capturing temporal dependencies in range-averaged currents (RACs), a multilayer perceptron (MLP) for nonlinear mapping to spatial Fourier coefficients, and a physics-constrained (PC) module that imposes a horizontal incompressibility constraint to enhance the accuracy of the predicted 3-D flow fields. Experimental results on real CAT data demonstrate that the proposed PC-LSTM-MLP framework effec-tively captures complex spatiotemporal dynamics and provides accurate predictions up to 25 minutes ahead. The model achieves an average mean absolute error (MAE) of 0.0241 m/s and an average root mean square error (RMSE) of 0.0269 m/s in 3-D flow field prediction. This research shifts CAT analysis from traditional inversion methods to intelligent, prediction-oriented modeling, underscoring the significant potential of deep learning in ocean flow field prediction.
- 20 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73 (00189456) , 1-15
Mirror-Type Signal Transmission Network Based on Acoustic Tomography for Continually Monitoring Flow Field Real-Time
Ocean current measurement via underwater acoustic tomography network communication is an advanced technology. The mirror-type coastal acoustic tomography (MCAT) technology is proposed to measure the current field with multi-station signal transmission. Here, the travel time is solved by mirror reciprocal transmission between station pairs based on three times communications. A four-station MCAT network experiment was initially performed to map the flow field variations near Liuheng island, ZhouShan, China. The multi-peak identification of the mirror transmitted signal is discussed and arrival peaks are matched via the ray simulation. And the hedging and mixing progresses of the two-dimensional current are reconstructed using the stream function inverse method. The mirror inversion results verified to be close to original results, but still exist the difference caused by instantaneous flow at different transmission timing. Based on mirror transmission function, MCAT observation could obtain deep-sea environment data from distant offshore sound stations in real time. In general, this technology obtains real-time observation and measurement data by signal communication during the observation process. Compared to recycling or in-situ observation, we believe such multi-station MCAT network provide long series monitoring of some dynamic processes like internal waves, mesoscale eddies, thermocline variations and so on.
- 20 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73 (00189456) , 1-15
Mirror-Type Signal Transmission Network Based on Acoustic Tomography for Continually Monitoring Flow Field Real-Time
Ocean current measurement via underwater acoustic tomography network communication is an advanced technology. The mirror-type coastal acoustic tomography (MCAT) technology is proposed to measure the current field with multi-station signal transmission. Here, the travel time is solved by mirror reciprocal transmission between station pairs based on three times communications. A four-station MCAT network experiment was initially performed to map the flow field variations near Liuheng island, ZhouShan, China. The multi-peak identification of the mirror transmitted signal is discussed and arrival peaks are matched via the ray simulation. And the hedging and mixing progresses of the two-dimensional current are reconstructed using the stream function inverse method. The mirror inversion results verified to be close to original results, but still exist the difference caused by instantaneous flow at different transmission timing. Based on mirror transmission function, MCAT observation could obtain deep-sea environment data from distant offshore sound stations in real time. In general, this technology obtains real-time observation and measurement data by signal communication during the observation process. Compared to recycling or in-situ observation, we believe such multi-station MCAT network provide long series monitoring of some dynamic processes like internal waves, mesoscale eddies, thermocline variations and so on.
- 8 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 21 (1545598X) , 1-5
3-D Ocean Temperature Prediction via Graph Neural Network With Optimized Attention Mechanisms
Ocean temperature prediction plays crucial roles in the ocean-related fields. The graph neural networks (GNNs) show advantages for modeling complex environmental issues. However, the prior methods typically focus on node features to predict ocean temperature. Note that the edge features are essential, especially for 3-D ocean temperature prediction. In this letter, a GNN with optimized attention mechanisms (GNN-OAMs) model is proposed to predict 3-D ocean temperature. The GNN-OAM introduces a random forest (RF) module to capture the nonstationary temporal dependencies. Especially, the OAMs, via combination of multiple adjacency matrices construct edge features, are presented to capture the dynamic spatial dependencies. The prediction performance of GNN-OAM model is evaluated in the 3-D ocean temperature experimental dataset, and the results show that the GNN-OAM achieves ocean temperature prediction with high accuracy, and the MAEs are 0.146 and 0.26, when predicting temperatures in next one day and five days.
- 8 May 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 21 (1545598X) , 1-5
3-D Ocean Temperature Prediction via Graph Neural Network With Optimized Attention Mechanisms
Ocean temperature prediction plays crucial roles in the ocean-related fields. The graph neural networks (GNNs) show advantages for modeling complex environmental issues. However, the prior methods typically focus on node features to predict ocean temperature. Note that the edge features are essential, especially for 3-D ocean temperature prediction. In this letter, a GNN with optimized attention mechanisms (GNN-OAMs) model is proposed to predict 3-D ocean temperature. The GNN-OAM introduces a random forest (RF) module to capture the nonstationary temporal dependencies. Especially, the OAMs, via combination of multiple adjacency matrices construct edge features, are presented to capture the dynamic spatial dependencies. The prediction performance of GNN-OAM model is evaluated in the 3-D ocean temperature experimental dataset, and the results show that the GNN-OAM achieves ocean temperature prediction with high accuracy, and the MAEs are 0.146 and 0.26, when predicting temperatures in next one day and five days.
- Open Access
- 21 March 2024
- Published by MDPI AG in Remote Sensing
- Vol. 16 (6) , 1105
Monitoring Thermal Exchange of Hot Water Mass via Underwater Acoustic Tomography with Inversion and Optimization Method
Thermal exchange of underwater water mass caused by marine heat wave is a hot point of research recently. In particular, because the water temperature observation along hot water mass transportation is hard work. Acoustic tomography is an advanced method to measure water temperature variations via sound signal transmission with multi-station network sensing. The 5 kHz frequency acoustic tomography used for observing water temperature variations caused by ocean heat waves is interesting work. In this paper, the numerical simulation of hot water mass is completed first, then floatation and diffusion of hot water mass in a simulation are monitored by acoustic tomography. A new inversion optimization method is proposed to obtain hot water mass transportation variations at two-dimensional temperature vertical profile. The proposed inversion method adds a regularized mode matrix and the optimization method adds the model correlation matrix to improve the results quality. The accuracy of inversion optimization results is compared and discussed, where the mean temperature error is less than 0.4 °C. Sensing water temperature variation of marine heat waves is verified via acoustic signal transmission and improved inversion optimization method. The water dynamical process observation is an application of acoustic tomography, which can be further used observe underwater environmental characteristics.
- 21 March 2024
- Published by MDPI AG in Remote Sensing
- Vol. 16 (6) , 1105
Open AccessMonitoring Thermal Exchange of Hot Water Mass via Underwater Acoustic Tomography with Inversion and Optimization Method
Thermal exchange of underwater water mass caused by marine heat wave is a hot point of research recently. In particular, because the water temperature observation along hot water mass transportation is hard work. Acoustic tomography is an advanced method to measure water temperature variations via sound signal transmission with multi-station network sensing. The 5 kHz frequency acoustic tomography used for observing water temperature variations caused by ocean heat waves is interesting work. In this paper, the numerical simulation of hot water mass is completed first, then floatation and diffusion of hot water mass in a simulation are monitored by acoustic tomography. A new inversion optimization method is proposed to obtain hot water mass transportation variations at two-dimensional temperature vertical profile. The proposed inversion method adds a regularized mode matrix and the optimization method adds the model correlation matrix to improve the results quality. The accuracy of inversion optimization results is compared and discussed, where the mean temperature error is less than 0.4 °C. Sensing water temperature variation of marine heat waves is verified via acoustic signal transmission and improved inversion optimization method. The water dynamical process observation is an application of acoustic tomography, which can be further used observe underwater environmental characteristics.
- Open Access
- 9 February 2024
- Published by MDPI AG in Remote Sensing
- Vol. 16 (4) , 646
Prediction of Water Temperature Based on Graph Neural Network in a Small-Scale Observation via Coastal Acoustic Tomography
Coastal acoustic tomography (CAT) is a remote sensing technique that utilizes acoustic methodologies to measure the dynamic characteristics of the ocean in expansive marine domains. This approach leverages the speed of sound propagation to derive vital ocean parameters such as temperature and salinity by inversely estimating the acoustic ray speed during its traversal through the aquatic medium. Concurrently, analyzing the speed of different acoustic waves in their round-trip propagation enables the inverse estimation of dynamic hydrographic features, including flow velocity and directional attributes. An accurate forecasting of inversion answers in CAT rapidly contributes to a comprehensive analysis of the evolving ocean environment and its inherent characteristics. Graph neural network (GNN) is a new network architecture with strong spatial modeling and extraordinary generalization. We proposed a novel method: employing GraphSAGE to predict inversion answers in OAT, using experimental datasets collected at the Huangcai Reservoir for prediction. The results show an average error 0.01% for sound speed prediction and 0.29% for temperature predictions along each station pairwise. This adequately fulfills the real-time and exigent requirements for practical deployment.
- 9 February 2024
- Published by MDPI AG in Remote Sensing
- Vol. 16 (4) , 646
Open AccessPrediction of Water Temperature Based on Graph Neural Network in a Small-Scale Observation via Coastal Acoustic Tomography
Coastal acoustic tomography (CAT) is a remote sensing technique that utilizes acoustic methodologies to measure the dynamic characteristics of the ocean in expansive marine domains. This approach leverages the speed of sound propagation to derive vital ocean parameters such as temperature and salinity by inversely estimating the acoustic ray speed during its traversal through the aquatic medium. Concurrently, analyzing the speed of different acoustic waves in their round-trip propagation enables the inverse estimation of dynamic hydrographic features, including flow velocity and directional attributes. An accurate forecasting of inversion answers in CAT rapidly contributes to a comprehensive analysis of the evolving ocean environment and its inherent characteristics. Graph neural network (GNN) is a new network architecture with strong spatial modeling and extraordinary generalization. We proposed a novel method: employing GraphSAGE to predict inversion answers in OAT, using experimental datasets collected at the Huangcai Reservoir for prediction. The results show an average error 0.01% for sound speed prediction and 0.29% for temperature predictions along each station pairwise. This adequately fulfills the real-time and exigent requirements for practical deployment.
- 20 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-18
3-D Water Temperature Distribution Observation via Coastal Acoustic Tomography Sensing Network
Water temperature is a critically important parameter for characterizing aquatic ecosystem, particularly in hydrodynamic processes produced by human activities or climate effects in shallow water regions (oceanic pasture, tidal currents, upwelling currents, etc). The variation of water temperature is a slight and slow process, as technology advances, high-precision observation for of the spatial and temporal variation of water temperature in three-dimensional scale is crucial to environment research and predictions. To address this issue, three-dimensional water temperature field is reconstructed by costal acoustic tomography technology, and inversion water temperature variations are obtained. A multi acoustic station experiment was conducted in a reservoir in Changsha, China, from March 4th to March 5th in 2022. Two grid-averaged three-dimensional inversion methods (Lagrangian-Least-Square method and Improved-Landweber iteration method) with multi-station underwater sensing network are processed for reveal water temperature fluctuations. Rotated empirical orthogonal function is used to further analyze and evaluate the sensitivity of temperature variations in the individual spatial grids. In this paper, water temperature distribution and dynamic variation process in the whole spatial structure is obtained and discussed using three-dimensional sound transmission information, and feasibility of the methods are verified by comparing with temperature depth sensor array observation results. Our study provides high-precision ideas and methods for three-dimensional water temperature activities of underwater acoustic tomography network in shallow environments.
- 20 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-18
3-D Water Temperature Distribution Observation via Coastal Acoustic Tomography Sensing Network
Water temperature is a critically important parameter for characterizing aquatic ecosystem, particularly in hydrodynamic processes produced by human activities or climate effects in shallow water regions (oceanic pasture, tidal currents, upwelling currents, etc). The variation of water temperature is a slight and slow process, as technology advances, high-precision observation for of the spatial and temporal variation of water temperature in three-dimensional scale is crucial to environment research and predictions. To address this issue, three-dimensional water temperature field is reconstructed by costal acoustic tomography technology, and inversion water temperature variations are obtained. A multi acoustic station experiment was conducted in a reservoir in Changsha, China, from March 4th to March 5th in 2022. Two grid-averaged three-dimensional inversion methods (Lagrangian-Least-Square method and Improved-Landweber iteration method) with multi-station underwater sensing network are processed for reveal water temperature fluctuations. Rotated empirical orthogonal function is used to further analyze and evaluate the sensitivity of temperature variations in the individual spatial grids. In this paper, water temperature distribution and dynamic variation process in the whole spatial structure is obtained and discussed using three-dimensional sound transmission information, and feasibility of the methods are verified by comparing with temperature depth sensor array observation results. Our study provides high-precision ideas and methods for three-dimensional water temperature activities of underwater acoustic tomography network in shallow environments.
- 27 June 2023
- Published by Elsevier in Ocean Engineering
- Vol. 284
Underactuated hovering control of the minitype autonomous underwater helicopter
- 27 June 2023
- Published by Elsevier in Ocean Engineering
- Vol. 284
Underactuated hovering control of the minitype autonomous underwater helicopter
- Open Access
- 7 April 2023
- Published by MDPI AG in Remote Sensing
- Vol. 15 (8) , 1965
Water Temperature Reconstruction via Station Position Correction Method Based on Coastal Acoustic Tomography Systems
Underwater acoustic tomography is an advanced technology in water environment observation. Sound propagation duration between transceivers is used for underwater parameter distribution profile reconstruction in the inverse problem. The key points of acoustic tomography are accurate station distance and time synchronization. Two methods are introduced in this study for sound station position correction. The direct signal transmission correction (DSC) method corrects the multi-peak (expect direct ray) travel time via the travel time difference between different sound rays and reference direct ray. The ray-model position correction (RMC) method calculates exact station position by the station drift distance obtained from transceiver depth variations to correct direct ray travel time; the other multi-peak travel time is revised by the corrected direct ray travel time. A water temperature observation experiment was carried out in a reservoir using coastal acoustic tomography (CAT) systems to verify the flexibility of these two methods. Multi-ray arrival peaks are corrected using DSC and RMC methods; water temperature inversion results in a 2D vertical profile are obtained. The reliability of the method is proved by comparison with temperature depth sensor (TD) data. The methods improve the quality of initial data and can be attempted for further water environment observation in acoustic tomography observation studies.
- 7 April 2023
- Published by MDPI AG in Remote Sensing
- Vol. 15 (8) , 1965
Open AccessWater Temperature Reconstruction via Station Position Correction Method Based on Coastal Acoustic Tomography Systems
Underwater acoustic tomography is an advanced technology in water environment observation. Sound propagation duration between transceivers is used for underwater parameter distribution profile reconstruction in the inverse problem. The key points of acoustic tomography are accurate station distance and time synchronization. Two methods are introduced in this study for sound station position correction. The direct signal transmission correction (DSC) method corrects the multi-peak (expect direct ray) travel time via the travel time difference between different sound rays and reference direct ray. The ray-model position correction (RMC) method calculates exact station position by the station drift distance obtained from transceiver depth variations to correct direct ray travel time; the other multi-peak travel time is revised by the corrected direct ray travel time. A water temperature observation experiment was carried out in a reservoir using coastal acoustic tomography (CAT) systems to verify the flexibility of these two methods. Multi-ray arrival peaks are corrected using DSC and RMC methods; water temperature inversion results in a 2D vertical profile are obtained. The reliability of the method is proved by comparison with temperature depth sensor (TD) data. The methods improve the quality of initial data and can be attempted for further water environment observation in acoustic tomography observation studies.
- 7 November 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 60 (01962892) , 1-14
Tomographic Mapping of Water Temperature and Current in a Reservoir by Trust-Region Method Based on CAT
Remote sensing of water parameter and flow dynamic progress is of critical importance to water environment observation. Underwater acoustic tomography technology uses multiple sound stations to construct distributed sensing network for large scale measuring, which has big superiority in water parameter observation compared with fix-point measurement. Water parameter variation progress can be monitored with underwater acoustic tomography sensing system. This paper develops a Grid-averaged Trust-Region Method for water parameter detailed profiling with coastal acoustic tomography (CAT) systems. A remote sensing experiment was conducted with 5 Coastal Acoustic Tomography (CAT) systems from February 28th to March 7th, 2022 in the Huangcai Reservoir in Changsha, China. Four CAT systems were deployed in same position with different depth of sound transceivers and another CAT station was set 650m apart. Water temperature and flow current reconstruction results along sound propagate section was analyzed with trust-region method and compared with the temperature data acquired with temperature sensor. The accuracy and feasibility of this method are verified with observation results.
- 7 November 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 60 (01962892) , 1-14
Tomographic Mapping of Water Temperature and Current in a Reservoir by Trust-Region Method Based on CAT
Remote sensing of water parameter and flow dynamic progress is of critical importance to water environment observation. Underwater acoustic tomography technology uses multiple sound stations to construct distributed sensing network for large scale measuring, which has big superiority in water parameter observation compared with fix-point measurement. Water parameter variation progress can be monitored with underwater acoustic tomography sensing system. This paper develops a Grid-averaged Trust-Region Method for water parameter detailed profiling with coastal acoustic tomography (CAT) systems. A remote sensing experiment was conducted with 5 Coastal Acoustic Tomography (CAT) systems from February 28th to March 7th, 2022 in the Huangcai Reservoir in Changsha, China. Four CAT systems were deployed in same position with different depth of sound transceivers and another CAT station was set 650m apart. Water temperature and flow current reconstruction results along sound propagate section was analyzed with trust-region method and compared with the temperature data acquired with temperature sensor. The accuracy and feasibility of this method are verified with observation results.
- Open Access
Layer-Averaged Water Temperature Sensing in a Lake by Acoustic Tomography with a Focus on the Inversion Stratification Mechanism
Continuous sensing of water parameters is of great importance to fluid dynamic progress study in oceans, coastal areas and inland waters. The acoustic tomography technique can perform water temperature field measurements horizontally and vertically using sound wave travel information. The layer-averaged water temperature can also be measured with the acoustic tomography method. However, investigations focusing on the stratified mechanism, which consists of stratification form and its influence on inversion error, are seldom performed. In this study, an acoustic tomography experiment was carried out in a reservoir along two vertical slices to observe the layer-averaged water temperature. Specifically, multi-path sound travel information is identified through ray tracing using high-precision topography data obtained via a ship-mounted ADCP during the experiment. Vertical slices between sound stations are divided into different layers to study layer division inversion methods in different preset types. The inversion method is used to calculate the average water temperature and inversion temperature error of every layer. Different layer methods are studied with a comparison of results. The layer division principle studied in this paper can be used for layer-averaged water temperature sensing with multi-path sound transmission information.
Open AccessLayer-Averaged Water Temperature Sensing in a Lake by Acoustic Tomography with a Focus on the Inversion Stratification Mechanism
Continuous sensing of water parameters is of great importance to fluid dynamic progress study in oceans, coastal areas and inland waters. The acoustic tomography technique can perform water temperature field measurements horizontally and vertically using sound wave travel information. The layer-averaged water temperature can also be measured with the acoustic tomography method. However, investigations focusing on the stratified mechanism, which consists of stratification form and its influence on inversion error, are seldom performed. In this study, an acoustic tomography experiment was carried out in a reservoir along two vertical slices to observe the layer-averaged water temperature. Specifically, multi-path sound travel information is identified through ray tracing using high-precision topography data obtained via a ship-mounted ADCP during the experiment. Vertical slices between sound stations are divided into different layers to study layer division inversion methods in different preset types. The inversion method is used to calculate the average water temperature and inversion temperature error of every layer. Different layer methods are studied with a comparison of results. The layer division principle studied in this paper can be used for layer-averaged water temperature sensing with multi-path sound transmission information.