A5101747052- Underwater Acoustics Research
- Underwater Vehicles and Communication Systems
- Marine animal studies overview
- Geophysical Methods and Applications
- Speech and Audio Processing
- Oceanographic and Atmospheric Processes
- Geological and Geochemical Analysis
- Geochemistry and Geologic Mapping
- Seismic Waves and Analysis
- High-pressure geophysics and materials
- Blind Source Separation Techniques
- earthquake and tectonic studies
- Arctic and Antarctic ice dynamics
- Ultrasonics and Acoustic Wave Propagation
- Ocean Waves and Remote Sensing
- Geological and Geophysical Studies
- Structural Health Monitoring Techniques
- Geochemistry and Geochronology of Asian Mineral Deposits
- Geochemistry and Elemental Analysis
- Seismic Imaging and Inversion Techniques
- Methane Hydrates and Related Phenomena
- Maritime and Coastal Archaeology
- Acoustic Wave Phenomena Research
- Advanced Neural Network Applications
- Coastal and Marine Dynamics
China Tobacco
2025
Sun Yat-sen University
2007-2025
Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou)
2023-2025
Shanghai University
2024-2025
Guilin University of Technology
2016-2024
Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai)
2023-2024
Shenzhen Technology University
2024
Shenzhen University
2024
Chinese Academy of Sciences
2014-2023
Ningbo Institute of Industrial Technology
2023
A deep learning approach based on big data is proposed to locate broadband acoustic sources using a single hydrophone in ocean waveguides with uncertain bottom parameters. Several 50-layer residual neural networks, trained huge number of sound field replicas generated by an propagation model, are used handle the uncertainty source localization. two-step training strategy presented improve models. First, range discretized coarse (5 km) grid. Subsequently, within selected interval and depth...
A multi-task learning (MTL) method with adaptively weighted losses applied to a convolutional neural network (CNN) is proposed estimate the range and depth of an acoustic source in deep ocean. The input normalized sample covariance matrices broadband data received by vertical line array. To handle environmental uncertainty, both training validation are generated propagation model based on multiple possible sets parameters. sensitivity analysis investigated examine effect mismatched...
High iodine loading and high-temperature adaptability of the cathode are prerequisites to achieving high energy density at full battery level promoting practical application for zinc-iodine (Zn-I
The horizontal wavenumbers and modal depth functions are estimated by block sparse Bayesian learning (BSBL) for broadband signals received a vertical line array in shallow-water waveguides. dictionary matrix consists of multi-frequency derived from shooting methods given large set hypothetical wavenumbers. dispersion relation is also taken into account to generate the dictionary. In this dictionary, only few entries used describe pressure field. These represent associated With constraint...
This research aims to explore the application of deep learning in autonomous driving computer vision technology and its impact on improving system performance. By using advanced technologies such as convolutional neural networks (CNN), multi-task joint methods, reinforcement learning, this article analyzes detail image recognition, real-time target tracking classification, environment perception decision support, path planning navigation. Application process key areas. Research results show...
The dispersion and attenuation characteristics of Arctic Ocean guided waves have significant applications in the analysis under-ice acoustic fields inversion parameters ice, water, sediment, seafloor. However, determining complex relationship propagating a damped medium presents challenges. In this study, was modeled as viscoelastic ice-water-sediment-seafloor coupled system. Moreover, spectral method introduced to investigate waves. wave equations boundary conditions were discretized,...
The frequency-difference (FD) method uses the FD Hadamard product, comprising auto-products to model below-band acoustic fields and unintended cross-products, for efficient direction-of-arrival (DOA) estimation under spatial aliasing. Despite improved resolution from compressive sensing, spurious peaks arise as a result of cross-products lacking counterparts in sensing matrix. proposed addresses this by reconstructing matrix with full product applying sparse Bayesian learning estimate...
Using deep convolutional neural networks as primary learners and a network meta-learner, source ranging is solved regression problem with the ensemble learning method. Simulated acoustic data from propagation model are used training data. Real an experiment in South China Sea test to demonstrate performance. The results indicate that direct zone of water, signals received by very receiver can be estimate range underwater sound source. Within 30 km, mean absolute error predictions 1.0 km...
This paper proposes the use of gated feedback recurrent unit network (GFGRU), a learning-based sparse estimation algorithm, for multiple source localization in direct arrival zone deep ocean. The GFGRU, trained on sound field replicas single generated by an acoustic propagation model, is used to estimate ranges and depths sources without knowing number sources. performance GFGRU compared Bartlett processor, feedforward neural (FNN), Bayesian Learning (SBL) algorithm. Simulations indicate...
A feature matching method based on the convolutional neural network (named FM-CNN), inspired from matched-field processing (MFP), is proposed to estimate source depth in shallow water. The FM-CNN, trained acoustic field replicas of a single generated by an propagation model range-independent environment, used and multiple depths mildly range-dependent environments. performance FM-CNN compared conventional MFP method. Sensitivity analysis for two methods performed study impact different...
The nonsynchronous measurements of microphone array are a powerful method for achieving large or high density by scanning the object from sequential movement prototype array. It has attracted great interest recently because it is beyond fundamental limitation working frequency that determined aperture and an A crucial problem to recover missing phase relation information between consecutive positions. in traditional solution generally attributed matrix completion block diagonal spectral...
The accurate prediction of sea surface temperature (SST) is essential for studying marine phenomena, understanding climate dynamics, and forecasting environmental changes. However, developing a general SST model challenging due to significant regional variations the impacts diverse phenomena. To improve performance predictions, we propose hybrid framework that effectively models spatial temporal dependencies data with Gaussian process-enhanced Long Short-Term Memory network. LSTM module...