A5101770885- Robotics and Sensor-Based Localization
- Control and Dynamics of Mobile Robots
- Advanced Image and Video Retrieval Techniques
- Robotic Locomotion and Control
- Industrial Vision Systems and Defect Detection
- Robotic Path Planning Algorithms
- Image Retrieval and Classification Techniques
- Manufacturing Process and Optimization
- Prosthetics and Rehabilitation Robotics
- Advanced Measurement and Metrology Techniques
- Indoor and Outdoor Localization Technologies
- Image and Object Detection Techniques
- 3D Surveying and Cultural Heritage
- Advanced Neural Network Applications
- Remote Sensing and LiDAR Applications
- Recommender Systems and Techniques
- Remote Sensing and Land Use
- Soil Mechanics and Vehicle Dynamics
- Image and Signal Denoising Methods
- Wireless Sensor Networks and IoT
- Optical measurement and interference techniques
- Text and Document Classification Technologies
- Surface Roughness and Optical Measurements
- Advanced Vision and Imaging
Yanshan University
2022-2024
Nanjing University of Science and Technology
2021
National University of Defense Technology
2006-2009
University of Hong Kong
2005
The most commonly used simultaneous localization and mapping (SLAM) scheme often assumes a static environment, leading to significant errors in pose estimation when operating highly dynamic scenes. To address this limitation improve the robustness accuracy of positioning environments, study proposes CD-SLAM, real-time stereo vision inertial SLAM system specifically designed for complex based on ORB-SLAM3. CD-SLAM enhances tracking thread introduces new parallel that utilizes YOLOv5 detect...
The semantic segmentation of LiDAR in the outdoor environment is still an open problem field automatic driving. Although emerging development and technological advancements, it remains challenging for three reasons: 1) uneven distribution Points 3D space; 2) same object can be represented by different point clouds sequences; 3) no unique geometric information compared with pixels 2D RGB images. In order to tackle all challenges mentioned above, we propose a new algorithm, Hybrid CNN-LSTM,...
Simultaneous localization and mapping (SLAM) has been widely used in augmented reality (AR), virtual (VR), robotics, autonomous vehicles as the theoretical basis for robots to perceive their environment. Most popular SLAM algorithms assume that objects scene are static. Solving dynamic problems is now attracting increasing attention. In this paper, we propose a method combines semantic segmentation information spatial motion of associated pixels cope with based on ORB-SLAM2. We add deep...
Abstract Simultaneous localization and mapping (SLAM) has been widely used in augmented reality(AR), virtual reality(VR), robotics, autonomous vehicles as the theoretical basis for robots to perceive their environment. Most popular SLAM algorithms assume that objects scene are static. Solving dynamic problems is now attracting increasing attention. In this paper, we propose a method combines semantic segmentation information spatial motion of associated pixels cope with algorithms. Our...
A novel stability criterion for underactuated biped robot is presented. The contacts ground with points and has not feet while walking running. definition of given. Loss Balance Degree(LBD) proposed as a according to the given before. LBD Linear Inverted Pendulum Mode (LIPM) computed specifically. entitled by analysing basic properties that any should has. conditions gait are concluded using LBD.
Visual Place Recognition (VPR) aims to robustly identify locations by leveraging image retrieval based on descriptors encoded from environmental images. However, drastic appearance changes of images captured different viewpoints at the same location pose incoherent supervision signals for descriptor learning, which severely hinder performance VPR. Previous work proposes classifying manually defined rules or ground truth labels viewpoints, followed training classification results. not all...
Abstract. Traditional indoor positioning technologies mostly require advanced installation of hardware devices, resulting in high costs and long-term maintenance. With advancements image recognition deep learning technologies, visual based on has become increasingly mature. This method offers the benefits low cost does not additional installation. However, it still inherent defects, such as cumbersome data collection, complex algorithms, universality. To minimize information pre-collection...
A novel biped robot and its control strategy are presented. The has cylindrical bar distributing in lateral plane as it's feet, which can acts a pivot when walking. Then the realize completely dynamic hybrid model is developed gait planned using of time-invariant. By finite-time nonlinear controller, realizes stable Simulation results show that limit cycle walking achieved, proposed feasible. Based on analysis relationship between velocity configuration, presented validated by simulation.
A new underactuated 3D biped robot and its control strategy are presented. Hybrid dynamics model is developed gait planned using the of time-invariant. By finite-time nonlinear controller, receives stable dynamic walking. Simulation results show that a limit cycle walking achieved, feasible. Based on analyses configuration velocity, method to velocity presented validated by simulation at last.
Abstract To ensure the reliability of carbon fiber material products and eliminate all kinds possible defects, it is necessary to take effective means check their quality. X-ray based nondestructive testing technology combined with image recognition algorithm considered be a fast efficient solution. However, surface processed often accompanied by labels which contains various information. These are confused in make most defect detection methods easy misidentify. This paper mainly studies...