A5066181483- Robotics and Sensor-Based Localization
- Robotic Path Planning Algorithms
- Robot Manipulation and Learning
- Video Surveillance and Tracking Methods
- Magnetic confinement fusion research
- Advanced Vision and Imaging
- Control and Dynamics of Mobile Robots
- Indoor and Outdoor Localization Technologies
- Human Pose and Action Recognition
- Anomaly Detection Techniques and Applications
- Robotic Locomotion and Control
- Fault Detection and Control Systems
- Target Tracking and Data Fusion in Sensor Networks
- Advanced Image and Video Retrieval Techniques
- Model Reduction and Neural Networks
- Advanced Control Systems Optimization
- Adaptive Control of Nonlinear Systems
- Autonomous Vehicle Technology and Safety
- Robotic Mechanisms and Dynamics
- Reinforcement Learning in Robotics
- Teleoperation and Haptic Systems
- Control Systems and Identification
- Neural Networks and Applications
- Image Enhancement Techniques
- Nuclear reactor physics and engineering
Guangdong Institute of Intelligent Manufacturing
2024
Dalian University of Technology
2023-2024
Guangdong Academy of Medical Sciences
2024
Southern Medical University
2024
Guangdong Provincial People's Hospital
2024
Wuyi University
2023-2024
Dalian University
2024
Shenzhen Institutes of Advanced Technology
2014-2023
Chinese Academy of Sciences
2014-2023
University of Chinese Academy of Sciences
2017-2023
Nowadays, the control technology of robotic manipulator with flexible joints (RMFJ) is not mature enough. The flexible-joint dynamic system possesses many uncertainties, which brings a great challenge to controller design. This paper motivated by this problem. In order deal and enhance robustness, full-state feedback neural network (NN) proposed. Moreover, output constraints RMFJ are achieved, improve security robot. Through Lyapunov stability analysis, we identify that proposed can...
This paper introduces a dynamic model of the driver control strategy lane-change behavior and applies it to trajectory planning in driver-assistance systems. The proposed reflects strategies adjusting longitudinal latitudinal acceleration during process can represent different driving styles (such as slow careful, well sudden aggressive) by using parameters. We also analyze features present methods for computing maximum position arrival time. Furthermore, we put forward an extended evasive...
Gesture recognition can be applied to many research areas, such as vision-based interface, communication and human robot interaction (HRI). This paper implements a non-intrusive, real-time gesture system using depth sensor. Related features are obtained from the skeleton model generated by Kinect Hidden Markov Models (HMMs) used dynamics of gestures. We conducted offline experiments check accuracy robustness system. Online were also performed verify requirement. Final results indicate that...
The approach of dynamical system (DS) is promising for modeling robot motion, and provides a flexible means realizing learning control. Accuracy, stability, speed are the three main factors to be considered when movements from human demonstrations with DS. Some approaches yield stable systems, but these may result in poor reproduction performance, while some good performance quite complex time-consuming. In this paper, we address accuracy-stability-speed issues simultaneously. We present...
The method of simultaneous localization and mapping (SLAM) using a light detection ranging (LiDAR) sensor is commonly adopted for robot navigation. However, consumer robots are price sensitive often have to use low-cost sensors. Due the poor performance LiDAR, error accumulates rapidly while SLAM, it may cause huge building larger map. To cope with this problem, paper proposes new graph optimization-based SLAM framework through combination LiDAR vision sensor. In framework, cost-function...
This paper focuses on finding the optimal sensor placement strategies for circular time-of-arrival (TOA) localization in three-dimensional (3D) space. The A-optimality criterion, minimizing trace of inverse Fisher information matrix (FIM), is applied to determine placements under Gaussian measurement noise. A comprehensive analysis sensor-target geometries provided general circumstance with no restriction number sensors, elevation angle and range. divided into two detailed sub-cases uniform...
Abstract Simultaneous localization and mapping systems based on rigid scene assumptions cannot achieve reliable positioning in a complex environment with many moving objects. To solve this problem, paper proposes novel dynamic multi-object lidar odometry (MLO) system semantic object recognition technology. The proposed enables the of robots objects generation long-term static maps scenes. For ego-motion estimation, extracts environmental features that take into account both geometric...
In a magnetic fusion reactor, the achievement of certain type plasma current profiles, which are compatible with magnetohydrodynamic stability at high pressure, is key to enable gain and non-inductive sustainment for steady-state operation. The approach taken toward establishing such profiles DIII-D tokamak create desired profile during ramp-up early flattop phases. evolution in time related poloidal flux, modeled normalized cylindrical coordinates using partial differential equation usually...
The Simultaneous Localization and Mapping (SLAM) of robots plays an increasingly important role in the development mobile robots. two most commonly used methods are visual LIDAR methods. In design, RGB-D camera is sensitive to light, it has a narrow field vision. Besides, featureless environment, difficult extract features. confines above lead poor tracking practice, which causes failure mapping completely. This paper aims at part SLAM using 2d low-cost finish robust indoor by mode switch...
Setting up a suitable current spatial profile in tokamak plasmas has been demonstrated to be key condition for one possible advanced scenario with improved confinement and steady-state operation. Experiments at the DIII-D focus on creating desired during plasma ramp-up early flattop phases aim of maintaining this target subsequent discharge. The evolution time is related poloidal magnetic flux, which modeled normalized cylindrical coordinates using parabolic partial differential equation...
An integrated plasma profile control strategy, ARTAEMIS , is being developed for extrapolating present-day advanced tokamak (AT) scenarios to steady-state operation. The approach based on semi-empirical modelling and was initially explored JET (Moreau et al 2008 Nucl. Fusion 48 106001 ). This paper deals with the general applicability of this strategy simultaneous magnetic kinetic various tokamaks. determination device-specific, control-oriented models that are needed compute optimal...
In this paper, we propose a new approach for anomaly detection in video surveillance. This is based on nonparametric Bayesian regression model built upon Gaussian process priors. It establishes set of basic vectors describing motion patterns from low-level features via online clustering, and then constructs to approximate the distribution kernel space. We analyze different measure criterions derived compare their performances. To reduce false detections caused by crowd occlusion, utilize...
This letter investigates how to place the received-signal-strength (RSS) sensors improve static target localization accuracy in three-dimensional (3-D) space. By using A-optimality criterion, i.e., minimizing trace of inverse Fisher information matrix (FIM), a new optimal RSS sensor placement strategy is developed when can be placed freely 3-D The smallest reachable Cramér-Rao lower bound, FIM, derived with corresponding sensor-target geometries. Besides, resistor network method and special...
In this paper, we present a novel method to detect two typical abnormal activities: pedestrain gathering and running. The is based on the potential energy kinetic energy. Reliable estimation of crowd density distribution are firstly introduced into detection anomalies. Estimation obtained from image model. By building foreground histogram X Y axis respectively, probability can be obtained, then define Crowd Distribution Index (CDI) represent dispersion. used pedestrains gathering. determined...