A5051266803- Photovoltaic System Optimization Techniques
- Solar Radiation and Photovoltaics
- Energy Load and Power Forecasting
- Neural Networks and Applications
- Blind Source Separation Techniques
- Stability and Control of Uncertain Systems
- Industrial Vision Systems and Defect Detection
- Solar Thermal and Photovoltaic Systems
- Advanced Control Systems Optimization
- Grey System Theory Applications
- Fault Detection and Control Systems
- Control Systems and Identification
- Electric Power System Optimization
- Visual Attention and Saliency Detection
- Advanced Image Fusion Techniques
- Face and Expression Recognition
- Advanced Image and Video Retrieval Techniques
- Advanced Algorithms and Applications
- Human Pose and Action Recognition
- Machine Fault Diagnosis Techniques
- EEG and Brain-Computer Interfaces
- Advanced Neural Network Applications
- Emotion and Mood Recognition
- Infrared Target Detection Methodologies
- Adaptive Control of Nonlinear Systems
Southeast University
2016-2025
Ministry of Education of the People's Republic of China
2021-2024
City University of Hong Kong, Shenzhen Research Institute
2024
Bridge University
2022
Center Point
2022
Oceaneering International (United States)
2022
China Southern Power Grid (China)
2015-2021
Southeast University
2016-2021
Power Grid Corporation (India)
2018
Shandong Institute of Automation
2006-2010
Epileptic seizure prediction is one of the most used therapeutic adjuvant strategies for drug-resistant epilepsy. Conventional approaches commonly collect training and testing samples from same patient due to inter-individual variability. However, challenging problem domain shift between various subjects remains unsolved, resulting in a low conversion rate clinic. In this work, adaptation (DA)-based model proposed circumvent issue. The short-time Fourier transform (STFT) employed extract...
Epileptic seizure prediction is one of the most used therapeutic adjuvant strategies for drug-resistant epilepsy. Conventional methods usually adopt handcrafted features and manual parameter setting. The over-reliance on expertise specialists may lead to weak exploitation low popularization clinical application. This paper proposes a novel parameterless patient-specific method based Fourier Neural Network (FNN), where transform backpropagation learning are synthesized make predictor more...
Boosted by a strong solar power market, the electricity grid is exposed to risk under an increasing share of fluctuant power. To increase stability grid, accurate forecast needed evaluate such fluctuations. In terms forecast, irradiance key factor generation, which affected atmospheric conditions, including surface meteorological variables and column integrated variables. These involve multiple numerical time-series images. However, few studies have focused on processing method data types in...
The photovoltaic (PV) cell temperature strongly affects the performance and efficiency of entire PV module. Thus, accurate estimation plays an important role in health monitoring energy assessment systems. This article proposes a multi-state dynamic thermal model for modules, considering heat-transfer mechanisms between module its environments, as well layers. proposed is benchmarked against field measurements at Aalborg University, Denmark. results demonstrate effectiveness to characterize...
Partial shading (PS) is inevitable during the outdoor operation of photovoltaic (PV) arrays. If not addressed appropriately, PS events will significantly reduce output power and induce other severe issues, such as hot spotting. In addition, maintenance strategies should be designed according to specific patterns, which are different from measures for faults, example, short circuiting. this article, a novel diagnosis methodology proposed detect classify shadings in PV strings, only electrical...
Abstract Stereo matching is one of the most popular techniques in diverse applications such as localization, 3D scene reconstruction, and autonomous driving. In pipeline stereo matching, cost volume construction aggregation are crucial components. Previous methods contain massive redundant features 4D require a great deal convolutions to aggregate cost, resulting high memory requirements computational costs. Moreover, current neglect geometric information, leading lower accuracy texture-less...