A5111683010- Robot Manipulation and Learning
- Reinforcement Learning in Robotics
- Soft Robotics and Applications
- Robotics and Sensor-Based Localization
- Robotic Path Planning Algorithms
- Robotics and Automated Systems
- Gaze Tracking and Assistive Technology
- Hand Gesture Recognition Systems
- Advanced Image and Video Retrieval Techniques
- Mobile Health and mHealth Applications
- Robotic Locomotion and Control
- Adversarial Robustness in Machine Learning
- Optical measurement and interference techniques
- Human Pose and Action Recognition
- Muscle activation and electromyography studies
- 3D Surveying and Cultural Heritage
- Teleoperation and Haptic Systems
- Domain Adaptation and Few-Shot Learning
- Tactile and Sensory Interactions
- COVID-19 Digital Contact Tracing
- Privacy, Security, and Data Protection
- Modular Robots and Swarm Intelligence
- Adaptive Dynamic Programming Control
- Cognitive Computing and Networks
- Signaling Pathways in Disease
Chinese Academy of Sciences
2012-2025
Institute of Automation
2018-2025
Center for Excellence in Brain Science and Intelligence Technology
2023-2024
Shanghai Institutes for Biological Sciences
2023-2024
NARI Group (China)
2024
Colorado State University
2024
Chengdu Institute of Biology
2024
Leshan Normal University
2022
University of Victoria
2022
Shandong Institute of Automation
2013-2022
Engineered extracellular vesicles (EVs) are considered excellent delivery vehicles for a variety of therapeutic agents, including nucleic acids, proteins, drugs, and nanomaterials. Recently, several studies have indicated that clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated 9 (Cas9) delivered by EVs enable efficient DNA editing. However, an RNA editing tool is still unavailable. Here, signal peptide-optimized EVs-delivered guide (gRNA) CRISPR/CasRx...
Abstract This paper presents a building‐level post‐hazard functionality model for communities exposed to flood hazards including the interdependencies between population, buildings, and infrastructure. An existing portfolio of building archetypes is used physical different typologies within community with goal supporting resilience‐informed decision‐making. Specific fragility‐based curves were developed this quantify exceedance probability prescribed set states. While buildings significant...
We address the problem of visual place recognition with perceptual changes. The fundamental is generating robust image representations which are not only insensitive to environmental changes but also distinguishable different places. Taking advantage feature extraction ability Convolutional Neural Networks (CNNs), we further investigate how localize discriminative landmarks that positively contribute similarity measurement, such as buildings and vegetations. In particular, a Landmark...
This research work is related to develop a motion control system of intelligent wheelchair based on hand gesture recognition for those with physical accessibility problem. In this paper, the accelerations in three perpendicular directions are detected by MEMS accelerometer and transmitted PC via Bluetooth wireless protocol. An automatic segmentation algorithm developed identify individual gestures sequence Hidden Markov Model(HMM) used model training. After this, trained Bayes method...
In teleoperation, the operator is often required to command motion of remote robot and monitor its behavior. However, such an interaction demands a heavy workload from human when facing with complex tasks dynamic environments. this article, we propose shared control method assist in manipulation reduce improve efficiency. We adopt task-parameterized hidden semi-Markov model learn skill several demonstrations. utilize learned predict target given current observed robotic trajectory...
Compared to traditional data-driven learning methods, recently developed deep reinforcement (DRL) approaches can be employed train robot agents obtain control policies with appealing performance. However, for real-world robots through DRL is costly and cumbersome. A promising alternative in simulated environments transfer the learned scenarios. Unfortunately, due reality gap between environments, often cannot generalized well real world. Bridging still a challenging problem. In this paper,...
In this paper, to better represent an athlete's performance using momentum, paper proposed indicator system assess a player's at different stages of match. The article tested both LightGBM and XGBoost classification models. By comparing the accuracy, recall, precision training sets for models, found that achieved 91.8% all these metrics, while impressive 98.4%. Therefore, used feature importance derived from model determine weight indicators. After obtaining weighted indicators players,...
Adapting the mastered manipulation skill to novel objects is still challenging for robots. Recent works have attempted endow robot with ability adapt unseen tasks by leveraging meta-learning. However, these methods are data-hungry in training phase, which limits their application real world. In this paper, we propose Meta-Residual Policy Learning (MRPL) reduce cost of policy learning and adaptation. During meta-training, MRPL accelerates process focusing on residual task-shared knowledge...
In this paper, the design of mechanism and control system a planetary wheeled stair-climbing wheelchair is introduced. The dynamic model wheel clusters established based on Lagrange equation, angle acceleration curves are studied with various given equivalent torques. Stability margin analyzed in detail during single step climbing procedure. According to simulation results, law wheelchair's derived from projection CG condition that stability always maintained. can be easily operated by an...
In unstructured environments, robotic grasping tasks are frequently required to interactively search for and retrieve specific objects from a cluttered workspace under the condition that only partial information about target is available, like images, text descriptions, 3D models, etc. It great challenge correctly recognize targets with limited learn synergies between different action primitives grasp densely occluding efficiently. this paper, we propose novel human-like push-grasping method...
This paper presents a novel facial movement based human machine interface for an intelligent wheelchair to operate in indoor environment. Five random selected intact subjects are designated control by using the proposed interface. The testing is carried out designed system performance evaluated implementing same tasks ten times, with criteria of easiness and time spent on each task. experimental result shows that this new strategy can assist disabled elderly users hands-free electric powered...
We present an approach to learn 3D local descriptor by combining both 2D texture and geometric information, which can be used register partial data for a variety of vision applications. Unlike previous approaches simply concatenate features learned from multiple sources into one feature descriptor, we representations jointly. design network, 3DTNet with architecture particularly designed learning robust representation leveraging information. Two types information are interacted each other...
Abstract In this paper, a novel self-adaptive underactuated robot hand with rigid-flexible coupling fingers (SAU-RFC hand) is proposed. The seven degrees of freedom (DOFs) SAU-RFC driven by four servomotors, consists three fingers, including two side-turning (ST) and one non-side-turning finger. Specially, the ST can perform synchronous reverse rotation laterally each other. Each finger joints DOFs introduces flexible structure, inner part proximal phalanx that makes most contact object...
Abstract “Picking out the impurities” is a typical scenario in production line which both time consuming and laborious. In this article, we propose target-oriented robotic push-grasping system able to actively discover pick impurities dense environments with synergies between pushing grasping actions. First, an attention module, includes target saliency detection density-based occluded-region inference. Without necessity of expensive labeling semantic segmentation, our module can quickly...
Robotic grasping in unstructured dense clutter remains a challenging task and has always been key research direction the field of robotics. In this paper, we propose novel robotic system that could use synergies between pushing actions to automatically grasp objects clutter. Our method involves using fully convolutional action-value functions (FCAVF) map from visual observations two tables Q-learning framework. These value infer utility actions, highest with corresponding location...
In the past decade, gait rehabilitation robot has drawn comprehensive attention. Robots greatly reduce workload of physical therapists and thus making sessions more effective widely available. Various research efforts have been made in field robot-assisted rehabilitation. One key concerns robotic assisted is planning customization. this paper, we first introduce a human measuring system based on binocular vision technique. Through system, data healthy subjects under different walking speed...
With the expansion of robotics application fields, robot always faces unpredictable manipulation tasks in hazardous and unstructured environments. Compared to tradition programming learning methods, teleoperation approaches shows great potential assisting operator perform accomplish complex uncertain tasks. In order achieve goal manipulating complicated with high accuracy, stability, convenience, this paper, we design a system based on virtual reality device. Our designed converts motion...
Aggregating neighbor features is essential for point cloud classification. In the existing work, each in may inevitably be selected as neighbors of multiple aggregation centers, all centers will gather from whole independently. Thus has to participate calculation repeatedly and generates redundant duplicates memory, leading intensive computation costs memory consumption. Meanwhile, pursue higher accuracy, previous methods often rely on a complex local aggregator extract fine geometric...