A5021924831- Robotics and Sensor-Based Localization
- Robotic Path Planning Algorithms
- Advanced Vision and Imaging
- Advanced Chemical Sensor Technologies
- Insect Pheromone Research and Control
- 3D Surveying and Cultural Heritage
- Olfactory and Sensory Function Studies
- Soft Robotics and Applications
- Autonomous Vehicle Technology and Safety
- Image and Object Detection Techniques
- Anomaly Detection Techniques and Applications
- Modular Robots and Swarm Intelligence
- Advanced Neural Network Applications
- Target Tracking and Data Fusion in Sensor Networks
- Control and Dynamics of Mobile Robots
- Advanced Image Processing Techniques
- Advanced Sensor and Control Systems
- Evacuation and Crowd Dynamics
- Geoscience and Mining Technology
- Reinforcement Learning in Robotics
- Rough Sets and Fuzzy Logic
- Dynamics and Control of Mechanical Systems
- Adaptive Control of Nonlinear Systems
- Generative Adversarial Networks and Image Synthesis
- Robot Manipulation and Learning
China University of Mining and Technology
2013-2024
Shanghai Jiao Tong University
2022
State Key Laboratory of Robotics and Systems
2022
Robotics Research (United States)
2022
Dalian Maritime University
2022
Chiba Institute of Technology
2022
Harbin Institute of Technology
2022
Chinese University of Hong Kong
2022
University of Nevada, Reno
2022
Beijing Academy of Artificial Intelligence
2021
Soft continuum robot is of superiority in the applications unstructured environments due to its high flexibility and safe interaction ability. Accurate shape control regarded as one prerequisites improve practicability. Considering situation where 3D position signals are not available, this letter investigates vision-based scheme a soft manipulator. The curvature features used depict reference image backbone curve solved through nonlinear optimization with given constraints two endpoints...
Soft robotics has attracted great attention because of its potential to overcome safety issues and produce a more harmonious cooperative environment for human beings robots. One the main challenges hindering wider application daily routine is dynamic modeling compliant mechanisms. In this paper, further investigate soft robot arm's performance extend utilization in water or other dense viscous mediums, model based on Kane's theory proposed. This takes complicated hydrodynamics into account...
Mobile robot autonomous navigation in large-scale environments with crowded dynamic objects and static obstacles is still an essential yet challenging task. Recent works have demonstrated the potential of using deep reinforcement learning to enable crowds. However, only considering human-robot interactions results short-sighted unsafe behaviors, they typically use hand-crafted features assume global observation range, leading large performance declines environments. advances shown power...
Visual servoing and vibration suppression of spatial flexible-link manipulators with a fixed camera setup are addressed in this article. The singular perturbation method is adopted to decouple the dynamic equations flexible manipulator; hence, two subsystems that represent rigid robot motion obtained, respectively. Then, for slow subsystem related motion, an image-based controller designed converge image errors consideration compensating approximating Jacobian matrix. For fast corresponding...
In this article, to solve trajectory tracing problem and vibration suppression for a planar two-link rigid-flexible manipulator subject joint-velocity measurement noise, novel uncalibrated visual servoing control is proposed. To begin with, the manipulator's dynamic model established by assumed mode method (AMM). On basis, based on singular perturbation theory, two subsystem controllers are designed, one slow controller, other fast controller. subsystem, cope with complication of camera...
Scene flow tracks the 3D motion of each point in adjacent clouds. It provides fundamental perception for autonomous driving and server robot. Although red green blue depth (RGBD) camera or light detection ranging (LiDAR) capture discrete points space, objects motions usually are continuous macroworld. That is, keep themselves consistent as they from current frame to next frame. Based on this insight, generative adversarial networks (GAN) is utilized self‐learn scene without ground truth. The...
3-D scene flow characterizes how the points at current time to next in Euclidean space, which possesses capacity infer autonomously nonrigid motion of all objects scene. The previous methods for estimating from images have limitations, split holistic nature by optical and disparity separately. Learning point clouds also faces difficulties gap between synthesized real data sparsity LiDAR clouds. In this article, generated dense depth map is utilized obtain explicit coordinates, achieves...
Glowworm swarm optimization (GSO) is a novel algorithm for the simultaneous computation of multiple optima multimodal functions, which intelligence based algorithm, such as ant colony (ACO) and particle (PSO). In multimode GSO performs very well in terms number peaks captured. this paper, we propose modified glowworm algorithm. Variable step-size movement strategy self-exploration behavior glowworms have been studied according to phenomena nature. way, accords with biological natural law...
As one of the important tasks in field robotics and machine vision, visual odometry provides tremendous help for various applications such as navigation, location, etc. Conventionally, task mainly relies on input continuous images. However, it is very complicated network to learn epipolar geometry information provided by Since 6 degree freedom (DoF) poses transformation occurs 3D space learning from point clouds more straightforward, this paper introduces concept pseudo-LiDAR task. The cloud...
Due to GPS restrictions, an inertial sensor is usually used estimate the location of indoor mobile robots. However, it difficult achieve high-accuracy localization and control by sensors alone. In this paper, a new method proposed robot pose with six degrees freedom based on improved 3D-Normal Distributions Transform algorithm (3D-NDT). First, point cloud data are captured Kinect segmented according distance robot. After segmentation, input processed Approximate Voxel Grid Filter in...
Based on introduction of the background and limitations present prediction methods for gas outburst in coal mines, this paper focuses introducing a new decision-making approach to with multi-sensor information fusion. Two fusion methods, neural network Dempster-Shafter evidence theory, were taken into account, improved combination rules D-S theory fuzzy sets was given decision Then practical experiment is prove efficiency effectiveness approach. The related experiments show that novel...
We will consider the problem of fast sampling control for singularly perturbed systems subject to actuator saturation and<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="M2"><mml:mrow><mml:msub><mml:mrow><mml:mi>L</mml:mi></mml:mrow><mml:mrow><mml:mn>2</mml:mn></mml:mrow></mml:msub></mml:mrow></mml:math>disturbance. A sufficient condition existence a state feedback controller is proposed. Under this controller, boundedness trajectories in presence of<mml:math...
The main purpose of traditional classification algorithms on bioinformatics application is to acquire better accuracy. However, these cannot meet the requirement that minimises average misclassification cost. In this paper, a new algorithm cost-sensitive regularised extreme learning machine (CS-RELM) was proposed by using probability estimation and cost reconstruct results. By improving accuracy group small sample which higher cost, CS-RELM can minimise 'rejection cost' integrated into...
This paper presents a novel method of simultaneously locating chemical sources by virtual physics-based multirobot system with release strategy. The proposed strategy includes setting forbidden area, releasing the robots from declared and escaping it rotary force goal force. can avoid relocating same source which has been located other leading them to move toward sources. Various turbulent plume environments are simulated Fluent Gambit software, set simulations performed on different...
In recent studies, the generalization of neural radiance fields for novel view synthesis task has been widely explored. However, existing methods are limited to objects and indoor scenes. this work, we extend outdoor scenes, trained only on object-level datasets. This approach presents two challenges. Firstly, significant distributional shift between training testing scenes leads black artifacts in rendering results. Secondly, viewpoint changes cause ghosting or missing regions rendered...
Estimation of the motion an agent and its environment concurrently is done by simultaneous localization mapping (SLAM). In recent past, SLAM has made rapid exciting progress used in different fields such as unmanned aerial vehicle (UAV), medical surgeries, endoscopic procedures. The aim this article to devise a more accurate physiotherapy exercise monitoring device on basis analysis from eight algorithms with criteria including power memory consumption, CPU heat, utilization. This provides...
This paper presents a virtual-physics force based control strategy for swarm robotic chemical source localization. The includes: structure formation force, goal and obstacle avoidant force. For formation, the robots maintain regular polygon virtual robot is located at center of polygon. motion depends on which obtained from sensor observations robots. Once moved to new place, would also move as single body with In this paper, we adopted chemotaxis plume tracing algorithms. Simulation...
Odor source detection of hazardous chemical leakage is significant for safety in production and life. On account its risk specificity, robot odor localization has been a consistent research scholars recent years. As the plume that was formed diffusion characteristics randomness intermittency, easy to be trapped into local optimum search process with low efficiency. The existing Fireworks Algorithm (FWA) improved by modifying explosion amplitude some extent mobile based on FWA (OSL-FWA) put...
Promising complementarity exists between the texture features of color images and geometric information LiDAR point clouds. However, there still present many challenges for efficient robust feature fusion in field 3D object detection. In this paper, first, unstructured clouds are filled 2D plane cloud extracted faster using projection-aware convolution layers. Further, corresponding indexes different sensor signals established advance data preprocessing, which enables cross-modal fusion. To...