A5057954028- Oceanographic and Atmospheric Processes
- Ocean Waves and Remote Sensing
- Marine and coastal ecosystems
- Geological and Geophysical Studies
- Underwater Acoustics Research
- Seismic Imaging and Inversion Techniques
- Robotic Path Planning Algorithms
- Tropical and Extratropical Cyclones Research
- Reinforcement Learning in Robotics
- Remote Sensing and Land Use
- Plant Physiology and Cultivation Studies
- Machine Fault Diagnosis Techniques
- Environmental Changes in China
- Methane Hydrates and Related Phenomena
- Distributed Control Multi-Agent Systems
- Wireless Signal Modulation Classification
- Visual Attention and Saliency Detection
- Advanced Image Fusion Techniques
- Model Reduction and Neural Networks
- Smart Grid Security and Resilience
- Mobile Crowdsensing and Crowdsourcing
- Cryospheric studies and observations
- Biometric Identification and Security
- Geophysics and Gravity Measurements
- Oil and Gas Production Techniques
Ocean University of China
2016-2025
Institute of Oceanology
2020-2024
Shenyang Center for Disease Control and Prevention
2024
Tsinghua University
2002-2024
Chinese Academy of Sciences
2014-2024
Tarim University
2023-2024
China Medical University
2024
University of Shanghai for Science and Technology
2024
Ministry of Natural Resources
2019-2022
Guangzhou Marine Geological Survey
2020-2022
In this paper, we propose a deep reinforcement learning (DRL)-based method that allows unmanned aerial vehicles (UAVs) to execute navigation tasks in large-scale complex environments. This technique is important for many applications such as goods delivery and remote surveillance. The problem formulated partially observable Markov decision process (POMDP) solved by novel online DRL algorithm designed based on two strictly proved policy gradient theorems within the actor-critic framework....
Unmanned Aerial Vehicles (UAVs) based delivery is thriving. In this paper, we model autonomous navigation of UAV in large-scale unknown complex environment as a discrete-time continuous control problem and solve it using deep reinforcement learning. Without path planning or map construction, our method enables UAVs to navigate from arbitrary departure places destinations only sensory information local GPS signal. We argue the task partially observable Markov decision process (POMDP) extant...
Internal waves (IWs) are a common characteristic of oceans and serve crucial role in transmitting energies between large-scale tides small-scale mixing. This study developed deep-learning-based method for extracting IW signatures on multiple satellite imagery from synthetic aperture radar (SAR) optical sensors sun-synchronous geostationary orbits with varying spatial resolution. We collected 1115 images, including 116 ENVISAT ASAR (Advanced SAR), 839 MODIS (MODerate-resolution Imaging...
Saline soil in seasonally frozen areas has caused tremendous damage to engineering and the ecological environment. The unfrozen water is main factor affecting properties of saline area therefore needs be studied. However, due high cost laboratory measurement content, this study focuses on using an adaptive network fuzzy inference system (ANFIS) a back propagation neural (BPNN) predict content Zhenlai area, Western Jilin. data for constructed model obtained by nuclear magnetic resonance (NMR)...
Internal waves (IWs), observed in the world oceans, have significant impacts on ocean engineering and environments. In this study, we collected satellite images from Moderate-Resolution Imaging Spectroradiometer Visible Infrared Radiometer Suite sensors Sulu-Celebes Sea 2016 to 2019 understand IW generation propagation. Satellite observations show a coherent phase difference both seas, indicating that IWs are alternatively generated when tidal currents oscillate back forth Sulu Archipelago,...
Internal waves (IWs) are broadly distributed globally and have significant impacts on offshore engineering underwater navigation. The prediction of IW propagation is a challenging task because the complex factors involved. In this study, machine-learning model was developed to predict in Andaman Sea. based back-propagation neural network trained by 1189 samples, including crest length peak-to-peak distance IWs, extracted from 123 Moderate-Resolution Imaging Spectroradiometer (MODIS) images...
Motors are one of the most critical components in industrial processes due to their reliability, low cost and robust performance. Motor failure will lead shutdown a whole production line cause great loss. Therefore, accurate, reliable effective motor fault diagnosis must be performed. Currently, motors has gained much attention guarantee safe operations. In this paper, novel method is proposed for three-phase asynchronous using Long Short-Term Memory (LSTM) neural network, which possesses...
Abstract Fault diagnosis based on deep learning has become a hot research topic because of the successful application in other fields. Due to variable operating conditions and harsh environment, it is extremely difficult effectively diagnose some typical faults diesel engines. When environmental factors change, performance models also unstable. In order solve these problems, this paper proposes novel model, called multi-branch convolutional neural networks (MBCNNs) with an integrated...
We solve an important and challenging cooperative navigation control problem, Multiagent Navigation to Unassigned Multiple targets (MNUM) in unknown environments with minimal time without collision. Conventional methods are based on multiagent path planning that requires building environment map expensive real-time computations. In this article, we formulate MNUM as a stochastic game devise novel deep reinforcement learning (MADRL) algorithm learn end-to-end solution, which directly maps raw...
The amplitude of internal waves is an important parameter for the remote-sensing detection. However, there no analytic method developed inversion based on optical images. In this article, deep-learning model introduced to inverse a large number peak-to-peak distance and 15 types texture characteristic parameters images are computed, relationship between also investigated. addition, correlation difference these amplitude, three established by selecting different as input variables. Results...
This paper proposes a new radar signal classification algorithm based on auto-correlation function (ACF) and directed graphical model (DGM). The ACFs of analytic signals are calculated to magnify the discrimination different categories. A simple de-noising approach is introduced purify ACFs. Four features extracted from purified ACF. DGM used represent joint probability distribution four along with category classify unknown signals. Simulation results show effectiveness this algorithm.
Based on ERA-Interim reanalysis wave field data for the 36 years from 1979 to 2014, temporal and spatial distributions development potential of energy are studied in detail offshore relatively nearshore waters adjacent Zhoushan Islands. The results show that areas high located east southeast (water depths 10 m 65 m) is higher especially city Taizhou, Zhejiang Province, which suitable locations development. conclusions provide scientific guidance sea
Understanding the mechanism of different imaging characteristics is necessary for image interpretation and information extraction internal solitary waves (ISWs). In this article, experimental method used to reveal bright dark patterns in optical remote-sensing images. It provides a scientific interpret ISWs The results prove that there are two critical angles which important interpretation. angle related zenith light source, sensor, wave slope modulated by ISWs. When source fixed sensor...
Abstract. Internal waves (IWs) are an important ocean process in transmitting energy between multiscale dynamics, making them a crucial oceanic phenomenon. The South China Sea (SCS) is renowned for its frequent large-amplitude IW activities, emphasizing the importance of collecting and analyzing extensive observational data. In this study, we present comprehensive dataset covering northern SCS 112.40–121.32° E 18.32–23.19° N, spanning from 2000 to 2022 with 250 m spatial resolution....
Mobile crowd sensing (MCS) has attracted extensive attention as a promising method for environmental and data collection. However, due to the increasing computational tasks, bandwidth, computing pressure, traditional "Cloud-Edge–End" MCS is insufficient efficiently offload in real-time aggregation. Therefore, we consider novel framework based on "Cloud-Enhanced-Edge–End," uses idle users edge nodes (ENs) assist servers enhance power reduce delay energy consumption. To achieve efficient...