A5100438821- Simulation and Modeling Applications
- Blockchain Technology Applications and Security
- Robotic Path Planning Algorithms
- Distributed Control Multi-Agent Systems
- Machine Fault Diagnosis Techniques
- Software-Defined Networks and 5G
- Neural Networks Stability and Synchronization
- Quantum optics and atomic interactions
- Advanced Memory and Neural Computing
- IoT and Edge/Fog Computing
- Image Enhancement Techniques
- Advanced Wireless Communication Techniques
- Cloud Computing and Resource Management
- Advanced Wireless Network Optimization
- Fault Detection and Control Systems
- Laser-Matter Interactions and Applications
- stochastic dynamics and bifurcation
- Wireless Communication Networks Research
- Network Security and Intrusion Detection
- Industrial Technology and Control Systems
- Traffic Prediction and Management Techniques
- Vehicle License Plate Recognition
- Advanced Vision and Imaging
- Image and Object Detection Techniques
- Smart Grid Energy Management
Tsinghua University
2020-2025
Shenzhen Institutes of Advanced Technology
2024-2025
Chinese Academy of Sciences
2025
China University of Petroleum, East China
2024
Zhangzhou Normal University
2024
University of Freiburg
2024
Institute of Software
2024
Southern University of Science and Technology
2024
Beijing Institute of Technology
2024
Hefei University of Technology
2023
The discharge of Fukushima radioactively contaminated water into the Pacific Ocean started in August 2023, posing comprehensive threats to marine ecosystems and human health globally. This study introduces Contaminated Water Risk Factor (FCWRF), which integrates three components─radionuclide diffusion, bioaccumulation, global seafood trade─to evaluate spatiotemporal distribution risks based on actual practices. Results suggest that exceeding 2 orders magnitude beyond baseline will be...
A cluster of dc microgrids consisting multiple interconnected has great potential for improving the reliability and reducing cost power generation. In this paper, a distributed coordination control strategy is proposed to overcome series problems centralized control. This which adopts hierarchical structure based on droop method, then modified formula achieves multi-objective introduced voltage generation Besides, finite-time consensus algorithm applied obtain average value obtained in...
Fault diagnosis keeps an essential tool to ensure the safety and reliability of a motor system. Based on deep learning, fault models constructed by mining historical data equipment have received extensive attention. However, high computational cost constrains application learning for diagnosis, especially when coping with multisource data. Inspired model reduction neural network structure automatic search method, this article proposed light weight multisensory fusion induction diagnosis....
Aiming at the traditional single sensor vibration signal cannot fully express bearing running state, and in high noise background, algorithm is insufficient for fault feature extraction. This paper proposes a diagnosis based on multi-sensor hybrid multimodal fusion to achieve high-precision by leveraging operating state information of bearings high-noise environment fullest extent possible. First, horizontal vertical signals from two sensors are fused using principal component analysis,...
Fault location in a power distribution network is challenging task due to the presence of multilayer branches and short line lengths. Existing fault-location methods generally require measurements at both ends each branch, which requires large number measuring points. The placement points branch terminals an approach that can be used reduce Such point layout allows existing for networks determine fault after identifying faulted branches. However, these fail locate if cannot correctly...
For increasingly complex communication demands of large-scale AI systems, the Space-Air-Ground Integrated Network (SAGIN) better caters to but also raises concerns about resource scarcity and diversity. This paper innovatively combines Graph Pointer Neural Networks (GPNN) Reinforcement Learning (RL) enhance allocation efficiency. The method leverages advantages GPNN in handling graph data RL optimizing decisions dynamic environments. It targets optimization goal maximizing while minimizing...
This paper presents a smart thermostat based on ISO7730 standard's thermal comfort model-PMV-PPD model for improving energy efficiency and saving energy. In the thermostat, desirable level (PPD) set by consumers determines ON/OFF state of air-conditioner control signal is predicted mean vote (PMV), which calculated with indoor temperature, human factors other climate index precisely estimating people's level. Besides, potential applying demand response discussed theoretically installing...
Aiming at the path planning problem of unmanned aerial vehicle (UAV) base stations when performing search tasks, this paper proposes a Double DQN-state splitting Q network (DDQN-SSQN) algorithm that combines state and optimal to complete UAV based on Deep Reinforcement Learning DDQN algorithm. The method stores multidimensional information in categories uses targeted training obtain information. also references received signal strength indicator (RSSI) influence reward by agent, way reduces...
Tailored light-matter interactions in the strong coupling regime enable manipulation and control of quantum systems with up to unit efficiency
Clustering has long been an important research topic in machine learning, and is highly valuable many application tasks. In recent years, methods have achieved high clustering performance by applying deep generative models. this paper, we point out that directly using q(z|y, x) instead of resorting to the mean-field approximation (as adopted previous works) Gaussian Mixture Variational Auto-Encoder can benefit unsupervised task. We improve VAE, optimizing it with a Monte Carlo objective...
<abstract><p>The mean square synchronization for a class of general stochastic delayed neural networks is explored in this paper using pinning impulsive control (PIC). It evident that PIC combines the profits and control. Considering there time delay between allocation execution instructions practice, idea average (AID) brought to describe kind delay. Furthermore, actuality, with internal disturbance are more general. Accordingly, some appropriate criteria derived Lyapunov...
We propose a novel low-complexity multi-user detection algorithm based on tree search for unified signal model that applies to both direct sequence code-division multiple access (DS-CDMA) and multi-carrier CDMA (MC-CDMA) systems. Simulation results show the proposed outperforms most previously schemes, can achieve near-optimal (maximum likelihood) performance with very low complexity.
This paper improves the traditional algorithm for circuit diagram image recognition, which identifies each component as well topological relationship over an entire diagram. uses generalized nodes basis analysis of connection relationship, and compares recognition speed accuracy proposed method against printed handdrawn diagrams. The comparison results show that in this is superior to algorithm.
Electronic toll collection (ETC) is an important part of the intelligent transport system (ITS). This paper first introduces meaning ETC, then discusses composition and workflow, finally analyzes advantages disadvantages. Saturated headway calculated by a mathematical model formula to get capacity ETC lane. And through vissim simulation technology, we station about different lanes. Finally describe all level service provide reference for setting
The 5G network slicing technology can greatly improve the transmission efficiency of users and significantly reduce delay. Network Functions Virtualization (NFV) provides a new solution for more flexible deployment efficient resource integration. allocation virtual resources is often realized through Service Function Chain (SFC). This paper studies problem under slicing. By designing an SFC model VFN backup method, with high reliability intelligent offload realized.
Tailored light-matter interactions in the strong coupling regime enable manipulation and control of quantum systems with up to unit efficiency, applications ranging from information photochemistry. While are readily induced at valence electron level using long-wavelength radiation, comparable phenomena have been only recently observed short wavelengths, accessing highly-excited multi-electron inner-shell states. However, strong-field processes wavelengths has not possible, so far, due lack...
One of the main goals material design is to sift proper materials with properties we want. However, traditional method, synthesizing and testing each in laboratory, wastes time energy, actual want usually one a million which makes it more difficult. Here, develop generative framework give guidance on specific properties. Our mainly drove by several variants Generative Adversarial Networks (GANs) for data generation. trained 86 perovskite-type samples including their components information,...
In e-commerce platforms, coupons play a crucial role in boosting transactions. the customer-to-customer (C2C) marketplace, ensuring satisfaction of both buyers and sellers is essential. While buyer-focused marketing strategies often receive more attention, addressing needs equally important. Additionally, existing tend to optimize each promotion independently, resulting lack continuity between promotions unnecessary costs pursuit short-term impact within period. We introduce Dynamic...
Since the underwater complex environmental conditions seriously affect accuracy of target recognition, in view low traditional image recognition algorithms for and lack robustness extracting features, an enhancement method small targets is proposed. A scheme combined with YOLO-v4 improved algorithm. Firstly, K-means++ algorithm used to improve original input anchor box mechanism achieve accurate selection, data strategy based on geometric photometric transformation optimize training network....