A5100639939- Robotics and Sensor-Based Localization
- Advanced Image and Video Retrieval Techniques
- Robotic Path Planning Algorithms
- Robot Manipulation and Learning
- Advanced Vision and Imaging
- Robotic Mechanisms and Dynamics
- Advanced Neural Network Applications
- Video Surveillance and Tracking Methods
- Human Pose and Action Recognition
- Muscle activation and electromyography studies
- Adaptive Control of Nonlinear Systems
- Anomaly Detection Techniques and Applications
- 3D Surveying and Cultural Heritage
- Soft Robotics and Applications
- Hand Gesture Recognition Systems
- Face and Expression Recognition
- Prosthetics and Rehabilitation Robotics
- Multimodal Machine Learning Applications
- Robotics and Automated Systems
- Advanced Measurement and Metrology Techniques
- Teleoperation and Haptic Systems
- Iterative Learning Control Systems
- Emotion and Mood Recognition
- Motor Control and Adaptation
- Optical measurement and interference techniques
Harbin Institute of Technology
2016-2025
Central South University
2025
Wuhu Hit Robot Technology Research Institute
2023-2024
State Key Laboratory of Robotics and Systems
2014-2024
Beihang University
2017-2024
Fudan University
2024
Northeast Electric Power University
2023-2024
Tokyo Metropolitan University
2020-2024
Tianjin University
2024
State Grid Corporation of China (China)
2024
It has been established that the transfer of human adaptive impedance is great significance for physical human-robot interaction (pHRI). By processing electromyography (EMG) signals collected from muscles, limb could be extracted and transferred to robots. The existing interfaces rely only on visual feedback and, thus, may insufficient skill in a sophisticated environment. In this paper, haptic mechanism introduced result muscle activity would generate EMG natural manner, order achieve...
In this work, we introduce FARP-Net, an adaptive local-global feature aggregation and relation-aware proposal network for high-quality 3D object detection from pure point clouds. Our key insight is that learning irregular yet sparse cloud generating superb proposals are both pivotal detection. Technically, propose a novel layer (LGFAL) fully exploits the complementary correlation between local features global features, fuses their strengths adaptively via attention-based fusion module....
In this work, we observe that indoor 3D object detection across varied scene domains encompasses both universal attributes and specific features. Based on insight, propose SOFW, a synergistic optimization framework investigates the feasibility of optimizing tasks concurrently spanning several dataset domains. The core SOFW is identifying domain-shared parameters to encode attributes, while employing domain-specific delve into particularities each domain. Technically, introduce set...
This paper investigates the prescribed performance fixed-time tracking control problem for a class of second-order nonlinear systems with bounded disturbance and actuator saturation limit. In order to facilitate controller development, we introduce function output error transformation technique transform dynamics inequality constraints an equivalent unconstrained one. Different from existing work on guaranteeing control, incorporate sliding mode surface into design procedure cope...
In this work, we show that it is feasible to perform multiple tasks concurrently on point cloud with a straightforward yet effective multi-task network. Our framework, Poly-PC, tackles the inherent obstacles (e.g., different model architectures caused by task bias and conflicting gradients dataset domains, etc.) of learning cloud. Specifically, propose residual set abstraction (Res-SA) layer for efficient scaling in both width depth network, hence accommodating needs various tasks. We...
Knee exoskeletons are devices that can enhance users’ mobility and strength. Their compatibility with the user’s joint motion is crucial for proper functioning. Typically, knee hinges designed to replicate instantaneous center of rotation (ICR) an average knee. This study represents a exoskeleton worn by user as one DoF closed-kinematic chain, which enables calculating displacement between thigh. The problem optimizing yield low relative movements during flexion extension motions formulated...
Drug discovery is crucial for identifying candidate drugs various diseases.However, its low success rate often results in a scarcity of annotations, posing few-shot learning problem. Existing methods primarily focus on single-scale features, overlooking the hierarchical molecular structures that determine different properties. To address these issues, we introduce Universal Matching Networks (UniMatch), dual matching framework integrates explicit with implicit task-level via meta-learning,...
This brief investigates the fixed-time tracking problem for uncertain nonlinear systems with input saturation constraint under event-triggered scheme. An adaptive controller is proposed to drive error a residual set in fixed time, which can eliminate of "explosion complexity". Furthermore, order decrease communication burden between and actuator while maintaining system control performance, we co-design strategy, guarantees that injected into only when predefined event occurs. In addition,...
As scene coordinate regression (SCoRe) methods become prevailing in the area of visual camera localization, issue repetitive or sparse texture scenes continues to be a concern. Specifically, they will suffer from performance degeneration due ambiguous patterns caused by similarity. In this work, we propose novel network for localization through single RGB image, with our key insight that taking only high-level feature maps as input can difficult accurately model problem and utilizing rich...