A5010546586- Advanced Battery Technologies Research
- Electric and Hybrid Vehicle Technologies
- Supercapacitor Materials and Fabrication
- Electric Vehicles and Infrastructure
- Railway Systems and Energy Efficiency
- Microgrid Control and Optimization
- Advancements in Battery Materials
- Fault Detection and Control Systems
- Electrical Contact Performance and Analysis
- Railway Engineering and Dynamics
- Distributed Control Multi-Agent Systems
- Machine Fault Diagnosis Techniques
- IoT and Edge/Fog Computing
- Reliability and Maintenance Optimization
- Industrial Technology and Control Systems
- Cloud Computing and Resource Management
- Extremum Seeking Control Systems
- Caching and Content Delivery
- Generative Adversarial Networks and Image Synthesis
- Fuel Cells and Related Materials
- Adaptive Control of Nonlinear Systems
- Vehicle Dynamics and Control Systems
- Advanced Computational Techniques and Applications
- Neural Networks Stability and Synchronization
- Advanced Memory and Neural Computing
Central South University
2016-2025
Singapore Management University
2022-2023
Shanghai Ninth People's Hospital
2023
Shanghai Jiao Tong University
2023
University of Southampton
2023
Xinyu University
2022
ORCID
2018-2020
Institute of Computing Technology
2015
Chinese Academy of Sciences
2015
Renmin Hospital of Wuhan University
2007
State-of-the-art deep neural networks are still struggling to address the catastrophic forgetting problem in continual learning. In this paper, we propose one simple paradigm (named as S-Prompting) and two concrete approaches highly reduce degree of most typical learning scenarios, i.e., domain increment (DIL). The key idea is learn prompts independently across domains with pre-trained transformers, avoiding use exemplars that commonly appear conventional methods. This results a win-win game...
The concept of the Internet-of-Batteries (IoB) has recently emerged and offers great potential for control optimization battery utilization in electric vehicles (EV). This concept, which combines aspects Internet-of-Things (IoT) with latest advancements technology cloud computing, can provide a wealth new information about health performance. be used to improve management number ways, including continuous prognosis improved vehicle management. In this paper, we reviewed detail basic...
Typical layout-to-image synthesis (LIS) models generate images for a closed set of semantic classes, e.g., 182 common objects in COCO-Stuff. In this work, we explore the freestyle capability model, i.e., how far can it unseen semantics (e.g., attributes, and styles) onto given layout, call task Freestyle LIS (FLIS). Thanks to development large-scale pre-trained language-image models, number discriminative image classification object detection) trained on limited base classes are empowered...
Recent years have witnessed the flourish of new promising technologies to prolong lifetime wireless sensor networks (WSNs). Employing mobile vehicles for charging and data collection is able balance load WSNs, meanwhile, provide a reliable energy replenishment sustainable network operations. Different from traditional scheduling policies where nodes passively wait arrival vehicles, novel dynamic clustering based mobile-to-cluster (M2C) scheme proposed optimize service process both vehicle in...
This article studies the group coordinated control problem for distributed nonlinear multiagent systems (MASs) with unknown dynamics. Cloud computing are employed to divide agents into groups and establish networked multigroup-agent (ND-MGASs). To achieve coordination of all actively compensate communication network delays, a novel model-free adaptive predictive (NMFAPC) strategy combining theory method is proposed. In NMFAPC strategy, each agent described as time-varying data model, which...
As the two classical power allocation methods in battery-supercapacitor hybrid energy storage systems, split-frequency and power-level have been developed separately for many years. In this article, we made an attempt to integrate advantages of proposed adaptive frequency-split-based quantitative strategy. First, preallocation is quantitatively determined according state charge (SoC) batteries supercapacitors. Then, a windowed fast Fourier transform (FFT)-based spectrum calculation algorithm...
This paper focuses on the prevalent stage interference and performance imbalance of incremental learning. To avoid obvious learning bottlenecks, we propose a new framework, which leverages series stage-isolated classifiers to perform task at each stage, without from others. be concrete, aggregate multiple as uniform one impartially, first introduce temperature-controlled energy metric for indicating confidence score levels classifiers. We then an anchor-based self-normalization strategy...
The turbofan engine is a crucial component of the aircraft. In order to provide an appropriate maintenance for improve reliability system, it necessary estimate remaining useful life (RUL) engine. this paper, data-based RUL prediction method proposed using Light Gradient Boosting Machine (LightGBM). To capture more degradation information, time window row data and runtime are used as inputs after normalization. LightGBM works very well with these high-dimensional model easy interpret....
This report presents results from the Video Person Recognition Evaluation held in conjunction with 11th IEEE International Conference on Automatic Face and Gesture Recognition. Two experiments required algorithms to recognize people videos Point-and-Shoot Challenge Problem (PaSC). The first consisted of a tripod mounted high quality video camera. second contained acquired 5 different handheld cameras. There were 1401 each experiment 265 subjects. subjects, scenes, actions carried out by are...
Battery management systems (BMS) are crucial for electric vehicles because effective battery is essential to their safe and reliable operation. However, BMS experience some problems like high cost, reduced space, low efficiency, failure rates. This paper proposes an intelligent digital twin model the which utilized historical data obtained from real driving scenarios measure, estimate, predict, diagnose pack states. The consists of two parts, one state estimation based on regression second...
Electrically driven legged robots have become popular in recent years. However, the development of reliable energy supply systems and effective management strategies for with dramatically varying power requirements still needs to be explored. This article proposes a learning-based model predictive control (MPC) strategy battery–supercapacitor hybrid storage containing prediction unit an MPC adaptive weights. Firstly, mathematical robot is established, dual-layer long short-term memory...