A5100340717- Robotic Locomotion and Control
- Prosthetics and Rehabilitation Robotics
- Robot Manipulation and Learning
- Robotics and Sensor-Based Localization
- Advanced Vision and Imaging
- Manufacturing Process and Optimization
- Robotic Path Planning Algorithms
- Advanced Computational Techniques and Applications
- Advanced machining processes and optimization
- Advanced Image and Video Retrieval Techniques
- Robotic Mechanisms and Dynamics
- Higher Education and Teaching Methods
- Advanced Neural Network Applications
- Soft Robotics and Applications
- Reservoir Engineering and Simulation Methods
- Industrial Technology and Control Systems
- Indoor and Outdoor Localization Technologies
- Muscle activation and electromyography studies
- Robotics and Automated Systems
- Digital Transformation in Industry
- Spinal Cord Injury Research
- Image and Object Detection Techniques
- Educational Technology and Assessment
- Soil Mechanics and Vehicle Dynamics
- Flexible and Reconfigurable Manufacturing Systems
Beijing Institute of Technology
2015-2024
National Cheng Kung University
2005-2024
Zhengzhou Railway Vocational & Technical College
2023-2024
Tongji University
2006-2023
Ministry of Education of the People's Republic of China
2012-2023
Beijing Advanced Sciences and Innovation Center
2018-2023
Second Affiliated Hospital of Jilin University
2023
Beijing Forestry University
2023
Wenzhou Medical University
2023
ShanghaiTech University
2023
Long‐term continuous Global Positioning System (GPS) observations have become an important tool for studying the various geodynamic processes. To fully study processes at GPS stations, temporal movements of monuments and nearby bedrock induced by thermal expansion need to be considered. In this paper, we extend a theoretical model estimate expansions 86 globally distributed stations based upon measurements surface air temperatures. The results show that annual temperature variations are...
This paper presents the mechanical and control system design of latest humanoid robot platform, BHR-5, from Beijing Institute Technology. The was developed as a comprehensive platform to investigate planning for fast responsive motion under unforeseen circumstances, example, playing table-tennis. It has improvement on structure, stiffness, reliability. An open architecture based concurrent multichannel communication mode CAN bus is proposed upgrade real-time performance expansibility system....
Human-like features, like toe-off, heel-strike, and disturbance rejection, can enhance the performance of bipedal robots. However, required control strategies for these motions influence each other, few studies have considered them simultaneously. Humans walk stably with toe-off heel-strike even after experiencing disturbances. Thus, we study human strategies, then, apply to a robot. This paper proposes bioinspired method realize stable walking robot First, analyze obtain some strategies....
Humanoid robots are being designed to perform tasks currently carried out by human workers in industry, manufacturing, service, and disaster assistance. To this end, the humanoid robot should be able walk stably across many types of terrain. However, when traversing a complex unknown environment, it is difficult realize accurate terrain perception immediately through large data collected sensor system, leading difference between planned foot landing positions actual positions. As result, an...
Vibration signals acquired by accelerometers are commonly used in the field of bearing fault diagnosis. The measured vibration contain a series impulse responses when localized faults occur. These impulses often viewed as features and exhibit nonlocal self-similarity terms morphological characteristics. However, generally contaminated background noise. In this paper, group-based K-SVD denoising algorithm, which exploits property, is presented to extract features. contrast regular denoising,...
Leukemia diagnosis based on bone marrow cell morphology primarily relies the manual microscopy of smears. However, this method is greatly affected by subjective factors and tends to lead misdiagnosis. This study proposes using images employs convolutional neural network (CNN) combined with transfer learning establish an objective, rapid, accurate for classification LKA (AML, ALL, CML). We collected 104 smears (including 18 healthy subjects, 53 AML patients, 23 ALL CML patients). The perfect...
When a biped robot moves about in physical environment, it may encounter inclined ground. Biped walking on ground still remains challenging for robots. Previous studies have discussed along specific directions. However, omnidirectional walk has rarely been investigated. In this paper, we propose gait pattern generation method First, model that describes the motion of uniformly with two angle parameters is proposed. A mathematical relationship between motions sagittal and coronal planes are...
For real-world applications, a biped robot should maintain stable walking when subjected to sudden external disturbances. Typically, unexpected changes body acceleration indicate that is experiencing an disturbance. This paper presents controller for rapid response large First, novel adjustment algorithm foot placement proposed. Here, zero-moment point variations are mapped onto the new foothold based on calculations from in acceleration. Second, impact reduction control landing presented...
In online advertising, conventional post-click conversion rate (CVR) estimation models are trained using clicked samples. However, during serving the need to estimate for all impression ads, leading sample selection bias (SSB) issue. Intuitively, providing reliable supervision signals unclicked ads is a feasible way alleviate SSB This paper proposes an uncertainty-regularized knowledge distillation (UKD) framework debias CVR via distilling from ads. A teacher model learns click-adaptive...
5G and AI (Artificial Intelligence) are changing industrial production offer great potential for manufacturing enterprises. One of the effects resulting from increasing quantity data is demands transmission large amounts data, fast speed, rapid analysis. However, merely relying on traditional communication technology manual processing does not lead to high performance low analysis time. It essential integrate flexibly transmit real-time data. To demonstrate feasibility these two...
The manipulating robots receive much attention by offering better services, where object grasping is still challenging especially under background interferences. In this article, a novel two-stream convolutional neural network (CNN) with simultaneous detection and segmentation proposed. proposed method cascaded an improved BlitzNet CNN TsGNet. introduces the channel-based mechanism, achieves improvement of accuracy combination learning multitask loss weightings suppression. Based on obtained...
In the case of a single scene feature, positioning an indoor service robot takes long time, and localization errors are likely to occur. A new method for hybrid system according multi-sensor fusion is proposed solve these problems. The process divided in two stages: rough precise positioning. By virtue K nearest neighbors based on possibility (KNNBP) algorithm first created present study, position determined received signal strength indicator (RSSI) Wi-Fi. Then, particle filter (HPFL)...
This study investigated the feasibility of using fluorescence hyperspectral imaging technology to diagnose early‐stage gastric cancer. Fluorescence spectral images 76 patients who were pathologically diagnosed as non‐atrophic gastritis, premalignant lesions and cancer collected. spectra at 100‐pixel points randomly extracted after binarization. Diagnostic models constructed through partial‐least‐square discriminant analysis (PLS‐DA) support vector machine (SVM) algorithms. The prediction...
Since a biped robot tends to tip over easily due impacts, rough terrain or obstacle object, stable and reliable walking is very important achievement. Well designed foot plays crucial role in stabilization control for humanoid robot. In this paper, we present flexible design our robot, including landing impact absorption mechanism attitude estimation system. First, the mechanical of proposed. Rubber bushes rubber pads are adopted absorbing impacts. Then, implementation hardware software...
In this paper, a novel ball trajectory prediction method is proposed for robot ping-pong player. The based on the dynamic model in sky and bouncing between table. First, ambiguity of removed. Then, table to get relationship emergence speed incidence speed. corresponding coefficients models are trained according 3D Cartesian coordinates calculated from high stereo vision system. With known models, can be predicted if several initial points system given. Experiments show that able accurately...
A globally consistent map is the basis of indoor robot localization and navigation. However, built by Rao-Blackwellized Particle Filter (RBPF) doesn't have high global consistency which not suitable for long-term application in large scene. To address problem, we present an improved RBPF Lidar SLAM system with loop detection correction named LCPF. The efficiency accuracy depend on segmentation submaps. Instead dividing submap at fixed number laser scan like existing method, Dynamic Submap...
Abstract Variable stiffness actuators (VSAs) are widely explored as they could improve the safe performance for human-robot interaction and make system torque controllable based on internal compliance. This article presents a novel VSA cam-based relocation mechanism (CRM-VSA), which is utilized to change locations of pivot spring lever simultaneously. Consequently, such structure makes actuator compacted regulation designable, help engineers pursue different demands regulation. The...
The large flange parts in aero-engines are usually manually assembled. Collision damage caused by assembly alignment error often occurs the process, which affects accuracy and product reliability. machine vision-based methods achieve high-precision measurement of obtaining high-resolution images whole with a combination laser distance sensors. Hence, existing high costly inefficient. In this paper, new method based on principle coaxial for is proposed. proposed first obtains multiple image...