- Vibration Control and Rheological Fluids
- Hand Gesture Recognition Systems
- Vibration and Dynamic Analysis
- Human Pose and Action Recognition
- Robot Manipulation and Learning
- Magnetic Bearings and Levitation Dynamics
- Reinforcement Learning in Robotics
- Advanced Adaptive Filtering Techniques
- Cavitation Phenomena in Pumps
- Extenics and Innovation Methods
- Advanced Optical Network Technologies
- Gear and Bearing Dynamics Analysis
- Image and Video Stabilization
- Adaptive Dynamic Programming Control
- Rough Sets and Fuzzy Logic
- Industrial Technology and Control Systems
- Soft Robotics and Applications
- Laser and Thermal Forming Techniques
- Wireless Sensor Networks and IoT
- Control Systems and Identification
- Robotic Path Planning Algorithms
- Language, Discourse, Communication Strategies
- Face recognition and analysis
- Digital Marketing and Social Media
- Insurance and Financial Risk Management
Shanghai Jiao Tong University
2009-2022
Cornell University
2020
Beijing Haidian Hospital
2020
Wuhan University of Technology
2008-2018
Henan University of Technology
2014
Chinese Academy of Social Sciences
2013
Tongji University
2009-2012
Hsin Sheng College of Medical Care and Management
2005
In this paper, we present FingerTrak, a minimal-obtrusive wristband that enables continuous 3D finger tracking and hand pose estimation with four miniature thermal cameras mounted closely on form-fitting wristband. FingerTrak explores the feasibility of continuously reconstructing entire postures (20 joints positions) without needs seeing all fingers. We demonstrate our system is able to estimate posture by observing only outline hand, i.e., silhouettes from wrist using low-resolution (32 x...
In the framework of intelligent aided language learning, a real-time data-driven 3D visual pronunciation system Chinese IPA is proposed. First, high quality articulatory speech corpus including and data lips, tongue jaw movements collected through Electro-Magnetic Articulograph; second, modeling shape design motion synthesis conducted. The obtained by designing precise facial model internal external articulators. combining parameterized anatomical model. can thus illustrate slight...
Grasping in cluttered scenes is challenging for robot vision systems, as detection accuracy can be hindered by partial occlusion of objects. We adopt a reinforcement learning (RL) framework and 3D architectures to search feasible viewpoints grasping the use hand-mounted RGB-D cameras. To overcome disadvantages photo-realistic environment simulation, we propose large-scale dataset called Real Embodied Dataset (RED), which includes full-viewpoint real samples on upper hemisphere with amodal...
An adaptive control method with dynamic interpolation is proposed for the active longitudinal vibration of propulsion shafting systems. In such systems, dynamics change speed-dependent stiffness, which can result in a time-varying system as shaft speed changes time. A model established investigation interpolating (DIAM). this model, induced by disturbance exerted on propeller (the left mass) and force thrust bearing right mass), defines channel channel. The DIAM used to suppress transmission...
An adaptive controller for active vibration isolation is proposed on the basis of filtered- x least-mean-squares method and enhanced with an anti-saturation scheme. In this method, responses induced by a multi-tonal disturbance are decomposed into harmonic components in-parallel tracking filters attenuated independently. The centre frequencies adjusted recursively estimated signal components. scheme derived normalized algorithm, which adopted to prevent excessive adaptation weights in...
In this paper, we present BodyTrak, an intelligent sensing technology that can estimate full body poses on a wristband. It only requires one miniature RGB camera to capture the silhouettes, which are learned by customized deep learning model 3D positions of 14 joints arms, legs, torso, and head. We conducted user study with 9 participants in each participant performed 12 daily activities such as walking, sitting, or exercising, varying scenarios (wearing different clothes, outdoors/indoors)...
A novel model-free control method for actively suppressing tonal structural vibration is investigated through numerical simulation and experiment. The established on the least-mean-squares-based adaptive feedback principle, an instrumental mechanism adjusting controller parameters proposed to guarantee convergence of algorithm. In verification method, a shaft–plate system used derive lightly damped dynamic model, which obtained by frequency response function synthesis, while in experiment...
Investigated are modeling and control approaches for vibration analysis of two identical beams which coupled with fluid active mechanical links. In the beam system, orthogonal functions used to represent fluid-structure interaction is considered. Frequency Response Functions (FRFs) derived from governing equations superposition principle linear systems. beams, impulse response corresponding FRFs an adaptive algorithm employed attenuate transmission between beams. Natural frequencies, mode...
In visual servo control system, the design of loop is a key link to precise tracking. To improve tracking accuracy, we discuss FCMAC-based (fuzzy cerebellar model articulation controller) strategy. The controller includes basic proportional regulator and FCMAC controller. two blocks work together approximate visual-mapping through an adaptive online learning law adjust weights network. By mapping 2-D image characteristic space 3-D kinematic space, can achieve task turntable directly...
Hot pressure welding (HPW) is one of the inner lead bonding process which widely used in microelectronics manufacturing. This paper presents a practical digital temperature controller for HPW. Based on step response, an identification method developed to obtain nonlinear time-delay model heating-up control design. The input-output linearization problem solved by means second-order Pade approximation and static inverse transform, optimal obtained machine. experimental equipment designed based...
A novel control system design approach is proposed, based on backstepping theory and adaptive critic (ACD) for a class of uncertain nonlinear systems. It ensures the systems' stability accurately track reference commands under large parametric perturbation. This so that whole not divided into slow fast subsystems compared with dynamic inversion (DI). However, can ensure little robustness closed-loop To solve this problem, ACD employed, which designed without any information In ACD, variable...
Grasping in cluttered scenes is challenging for robot vision systems, as detection accuracy can be hindered by partial occlusion of objects. We adopt a reinforcement learning (RL) framework and 3D architectures to search feasible viewpoints grasping the use hand-mounted RGB-D cameras. To overcome disadvantages photo-realistic environment simulation, we propose large-scale dataset called Real Embodied Dataset (RED), which includes full-viewpoint real samples on upper hemisphere with amodal...
Abstract In order to ensure spacecraft qualification before the delivery, thermal tests should be performed with regard its performance specification. The aim of is examine function and under extreme vacuum conditions, also detect possible manufacturing defects or check rationality theoretical models[1]. Temperature control products one critical tasks in test, which has characteristics large inertia time-delay. Ordinary single method often leads poor even out control[2]. this paper, model...