A5101529136- Maritime Navigation and Safety
- Ship Hydrodynamics and Maneuverability
- Advanced Measurement and Detection Methods
- Advanced Image Fusion Techniques
- Adaptive Control of Nonlinear Systems
- Image Enhancement Techniques
- Advanced Image Processing Techniques
- Infrared Target Detection Methodologies
- Topic Modeling
- Vehicle Routing Optimization Methods
- Speech Recognition and Synthesis
- Advanced Sensor and Control Systems
- Domain Adaptation and Few-Shot Learning
- Cavitation Phenomena in Pumps
- Maritime Transport Emissions and Efficiency
- Advanced Algorithms and Applications
- Industrial Technology and Control Systems
- Model Reduction and Neural Networks
- CO2 Reduction Techniques and Catalysts
- Machine Fault Diagnosis Techniques
- Outsourcing and Supply Chain Management
- Scheduling and Timetabling Solutions
- Stock Market Forecasting Methods
- Remote Sensing and LiDAR Applications
- Robotic Path Planning Algorithms
University of Liverpool
2025
Wuhan Ship Development & Design Institute
2023-2024
East China University of Science and Technology
2024
East China Normal University
2023
Wuxi Taihu Hospital
2023
iQIYI (China)
2023
University of Chicago
2022
Toyota Technological Institute at Chicago
2022
Wuhan University of Technology
2015-2021
China Classification Society
2020
A traditional A-Star (A*) algorithm generates an optimal path by minimizing the cost. For a vessel, factors of length, obstacle collision risk, traffic separation rule and manoeuvrability restriction should be all taken into account for planning. Meanwhile, water current also plays important role in voyaging berthing vessels. In consideration these defects when it is used vessel planning, improved has been proposed. To specific, risk models obstacles (bridge pier, moored or anchored ship,...
Image quality assessment (IQA) represents a pivotal challenge in image-focused technologies, significantly influencing the advancement trajectory of image processing and computer vision. Recently, IQA has witnessed notable surge innovative research efforts, driven by emergence novel architectural paradigms sophisticated computational techniques. This survey delivers an extensive analysis contemporary methodologies, organized according to their application scenarios, serving as beneficial...
Abstract Since vessel dynamics could vary during maneuvering because of load changes, speed changing, environmental disturbances, aging mechanism, etc., the performance model‐based path following control may be degraded if controller uses same motion model all time. This article proposes an adaptive method based on least squares support vector machines (LS‐SVM) to deal with parameter changes model. The consists two components: online identification varying parameters and predictive (MPC)...
Low quality automatic identification system (AIS) data often mislead analysts to a misunderstanding of ship behavior analysis and making incorrect navigation risk assessments. It is therefore necessary accurately understand judge the problems in AIS before further behavior. Outliers were filtered existing methods based only on mathematical models where related are not utilized there lack visual exploration. Thus, human brain’s ability cannot be fully think visually for reasoning. In this...
To learn the optimal collision avoidance policy of merchant ships controlled by human experts, a finite-state Markov decision process model for ship is proposed based on analysis mechanism, and an inverse reinforcement learning (IRL) method cross entropy projection to obtain from expert’s demonstrations. Collision simulations in different encounters are conducted results show that obtained IRL has good inversion effect two kinds which indicate can effectively experts avoidance.
Rich image information is one of the important means through which unmanned surface vehicles effectively and reliably identify targets during autonomous navigation. However, adaptability traditional artificial design feature methods in target representation differentiation remains limite d due to diversity ship types, different scales, complex dynamic outdoor scenes. This study proposed a recognition method based on single shot multi-box detector (SSD) deep learning. First, training test...
The marine diesel engine propulsion system is a nonlinear with time delay. In order to realize the accurate and real-time control of speed, new method based on state-compensate extended state observer, backstepping beetle antennae search algorithm, that is, adaptive observer-backstepping, proposed. First all, response relationship model between speed fuel injection established basis mean value engine. Then, deal load disturbances parameter perturbation engine, observer used estimate lumped...
We present WeaverBird, an intelligent dialogue system designed specifically for the finance domain. Our harnesses a large language model of GPT architecture that has been tuned using extensive corpora finance-related text. As result, our possesses capability to understand complex financial queries, such as "How should I manage my investments during inflation?", and provide informed responses. Furthermore, incorporates local knowledge base search engine retrieve relevant information. The...
Building paired datasets in low-light enhancement entails significant cost and time, making such precious commodities. Many researchers focus on how to enable models learn more information from limited datasets. A prevalent strategy involves employing semi-supervised learning techniques enhance model performance through additional unpaired images. However, one of the main challenges faced is scarcity a vast number images same domain as original Consequently, we introduce image method using...
The unmanned surface vessel (USV) can provide effective means for ocean exploration and traffic regulation, has become a research hotspot in the field of marine shipping because its low cost, high efficiency other advantages. To realize accurate control USV movement technical guarantee realization autonomous navigation. course method modeling, using white noise simulation environment operation boat interference, MPC (Model Predictive Control) algorithm based on state space is used to design...
无人船自适应路径跟踪控制系统
Adam with decoupled weight decay, also known as AdamW, is widely acclaimed for its superior performance in language modeling tasks, surpassing $\ell_2$ regularization terms of generalization and optimization. However, this advantage not theoretically well-understood. One challenge here that though intuitively optimizes the regularized loss, it clear if AdamW a specific objective. In work, we make progress toward understanding benefit by showing implicitly performs constrained More...
Numerical Simulation methods of RANS and LES have been widely used to simulate the velocity field induced by propellers. In order combine advantages quick computation time from detailed information LES, this paper innovatively proposed a customized data-driven encoding-decoding network achieve mapping results results. Firstly, model was constructed extract features RANS/LES for propeller wake with additional concatenation during encoding steps. Then reconstructed original samples got...