A5100335821- Machine Learning and ELM
- Neural Networks and Applications
- Muscle activation and electromyography studies
- Domain Adaptation and Few-Shot Learning
- Prosthetics and Rehabilitation Robotics
- EEG and Brain-Computer Interfaces
- Advanced Memory and Neural Computing
- Energy Harvesting in Wireless Networks
- Sparse and Compressive Sensing Techniques
- Total Knee Arthroplasty Outcomes
- Water Quality Monitoring Technologies
- Optimization and Search Problems
- Energy Efficient Wireless Sensor Networks
- Parallel Computing and Optimization Techniques
- Image Enhancement Techniques
- Interconnection Networks and Systems
- Computer Graphics and Visualization Techniques
- Advanced Vision and Imaging
- Advanced Manufacturing and Logistics Optimization
- Advanced Data Storage Technologies
- Remote Sensing and Land Use
- Stroke Rehabilitation and Recovery
- Fault Detection and Control Systems
- Hydrological Forecasting Using AI
- Advanced Algorithms and Applications
Tianjin University
2025
Qingdao University of Science and Technology
2024
Beijing Technology and Business University
2022-2024
National Research Center for Rehabilitation Technical Aids
2016-2023
Tsinghua University
2020
National University of Defense Technology
2018
National Taiwan University
2017
University of Science and Technology Beijing
2015
Node localization is one of the promising research issues in Wireless Sensor Networks (WSNs). A novel node algorithm termed Kernel Extreme Learning Machines based on Hop-count Quantization (KELM-HQ) proposed. The proposed employs real number hop-counts between anchors and unknown nodes as training inputs locations targets for KELM training. method also test samples to compute by trained KELM. Simulation results demonstrate that KELM-HQ improves accuracy it outperforms state-of-the-arts methods.
Recovering the spatially-varying bidirectional reflectance distribution function (SVBRDF) from a single image in uncontrolled environments is challenging while essential for various applications. In this paper, we address highly ill-posed problem using convenient capture setup and carefully designed reconstruction framework. Our proposed setup, which incorporates an active extended light source mirror hemisphere, easy to implement even common users requires no calibration. These devices can...
Rehabilitation technologies based on brain-computer interface (BCI) have become a promising approach for patients with dyskinesia to regain movement. In BCI experiment, there is often necessary stage of calibration measurement before the feedback applications. To reduce time required initial training, it great importance method which can learn classify electroencephalogram (EEG) signals little amount training data. this paper, novel combination feature extraction and classification algorithm...
Prosthetic knee joint (PKJ) is an important apparatus for trans-femoral amputees to regain walking ability. This study has two objectives: (1) design a high performance and low-cost passive PKJ (2) evaluate the of PKJ. In proposed design, four-bar linkage was employed as mechanical structure, parallel spring damper were used connecting rods linkage. With spring, length rod variable buffer flexion angle can be generated, which consistent with that human joint. The regulate swing speed shank....
Heterogeneous CPU-GPU systems have recently emerged as an energy-efficient computing platform. A robust integrated simulator is essential to facilitate researches in this direction. While few simulators are available, similar tools that support OpenCL 2.0, a widely used new standard with promising heterogeneous features, currently missing. In paper, we extend the existing simulator, gem5-gpu, 2.0. addition, conduct experiments on extended see impact of features introduced by Our 2.0...
Synthetic aperture radar (SAR) image classification is a popular yet challenging research topic in the field of SAR interpretation. This paper presents new method based on extreme learning machine (ELM) and superpixel-guided composite kernels (SGCK). By introducing generalized likelihood ratio (GLR) similarity, modified simple linear iterative clustering (SLIC) algorithm firstly developed to generate superpixel for image. Instead using fixed-size region, shape-adaptive used exploit spatial...
Target detection is one of the most important tasks radar. Statistical signal processing methods for target have made several achievements; meanwhile, neural network-based also received increasing attention. The newly proposed broad learning system (BLS) an effective and efficient incremental without deep architecture, it has been successfully applied to classification problem. In this article, BLS radar detection. We design a detector using algorithm, it's very attractive that shows better...
The accuracy of thickness is an important standard to measure the strip quality. Therefore, it crucial accurately obtain a high precision thickness. ELM (extreme learning machine) based on clustering forecast method presented for hot rolled prediction. Firstly, strong correlation properties are obtained by data pretreatment, in order ensure effectiveness model. Then, analysis made about attribute data. Finally, network performed respectively each type This paper uses filed production...
Abstract-Nowadays, robot industry is booming, and the control requirements are higher with more user-friendly features. Therefore, it significant to study mechanism of human upper limb movement for optimization. In this study, a musculoskeletal model was established, motion data collected by stereo photogrammetry system, which used scaling match measurement physiological characteristics experimental subjects generated generalized coordinate values elbow shoulder joints. Then, Opensim...
Abstract Prediction is crucial to prevent the outbreaks of algal bloom. However, due time-varying and nonlinear characteristics bloom, which brings challenges for accurate prediction. Aiming at solving problem, we consider both prior information data-driven, propose an error compensation combination prediction model based on manifold regularized extreme learning machine fused shapelet (ECM-MRELM-FS). First, extracted as upward set discover predict value bloom typical evolution mode....