A5103189058- Robotic Path Planning Algorithms
- Robotics and Sensor-Based Localization
- Robotic Locomotion and Control
- Multimedia Communication and Technology
- Modular Robots and Swarm Intelligence
- Data Management and Algorithms
- Remote Sensing and LiDAR Applications
- Music and Audio Processing
- Advanced Algorithms and Applications
- Robotics and Automated Systems
- Video Analysis and Summarization
- Cryptographic Implementations and Security
- Image and Object Detection Techniques
- Wireless Sensor Networks and IoT
- Industrial Technology and Control Systems
- Rough Sets and Fuzzy Logic
- Chaos-based Image/Signal Encryption
- Advanced Manufacturing and Logistics Optimization
- Industrial Vision Systems and Defect Detection
- Underwater Vehicles and Communication Systems
- Remote Sensing in Agriculture
- Autonomous Vehicle Technology and Safety
- Advanced Steganography and Watermarking Techniques
- Advanced Image and Video Retrieval Techniques
- Inertial Sensor and Navigation
Wuhan University of Technology
2024-2025
Chongqing University of Posts and Telecommunications
2016-2024
University of Science and Technology Liaoning
2024
Qingdao University of Science and Technology
2010
Acquiring terrain information during robot locomotion is essential for autonomous navigation, gait selection, and trajectory planning. Quadruped robots, due to their biomimetic structures, demonstrate enhanced traversability over complex terrains compared other robotic platforms. Furthermore, the internal sensors of quadruped robots acquire rich terrain-related data across diverse terrains. This study investigates relationship between characteristics based on proprioception sensor data,...
3D multiobject tracking (MOT) is an important part of road condition detection and hazard warning algorithm in roadside systems autonomous driving systems. There a tricky problem MOT that the identity occluded object switches after it reappears. Given good performance 2D MOT, this paper proposes with deep learning based on algorithm. Firstly, detector was used to obtain oriented bounding boxes from point clouds. Secondly, Kalman filter for state estimation, reidentification match feature...
To provide an accurate surface defects inspection method and make the automation of robust image region interests(ROI) delineation strategy a reality in production line, multi-source CCD imaging based fuzzy-rough sets is proposed for hot slab quality assessment. The applicability presented devised system are mainly tied to strip, billet etcetera. In this work we take into account complementary advantages two common machine vision (MV) systems(line array traditional scanning (LS-imaging) area...
With the development of information age, use images for expression has become very common. In view security images, more and scholars begin to study image encryption. this paper, we briefly describe chaotic encryption field, concept chaos theory, characteristics discriminant criteria, common typical systems, analysis methods summary.
<title>Abstract</title> In outdoor environments, the odometer based on lidar sensors has accumulated errors, which may lead to map drift and construction failure. This paper proposes a SLAM method global descriptor loop detection. uses stable triangle (STD) descriptors for closure detection reduce cumulative errors. At same time, this also candidate loopback frame search algorithm inverse document frequency (IDF) ensure that high query efficiency can be maintained when number of increases....
Basketball court detection is the key step of video annotation, however present different color when using camera or in play court. In this paper adaptive Gaussian Mixture Model (GMM) used to build model basketball court, which annotation. The proposed algorithm extracts main court's GMM and IEM. Under these conditions annotated into in-play close-up. Experimental results show that algorithms are effective.
shots classification is the fundamental for sports video annotation. This paper uses non-supervised method to cluster into defined classes (in-play, close-up and free-throw) based on low-level features of image (the main color histogram). Specifically, we apply None Euclidean Relational Fuzzy C-means (NERFCM) after comparing four extracting frames. Experiments prove its efficiency sensitivity.
The main computational complexity of A* algorithm lies in the state detection nodes and selection with lowest cost. When scenario path planning is large, often occupies too much memory because searching many nodes, which seriously affects efficiency finding. To solve above problems, a new evaluation function used to find most suitable point on adjacent edge as search neighbor. Domain, can not only destination node faster but also plan shorter path. simulation results show that improved...
The path solved by A* algorithms will go through some unnecessary nodes and increase the length of path. In order to improve this drawback, when expanding nodes, location specific is calculated using evaluation function instead predefined fixed in grid map, so that direction can be expanded at any angle function. Simulation results show new algorithm get a shorter short time.
Positioning technology is the basis for realizing fully autonomous navigation of mobile robots. Particle filtering used as a non-parametric Bayesian filter to solve this problem. However, due particle degradation, random sampling error, large amount calculation, etc., positioning always has accuracy and efficiency problems. In order improve global initial laser, paper proposes algorithm fusion laser vision. The integrates filtering, beam model, visual observation model enable robot complete...