- Robot Manipulation and Learning
- Robotic Mechanisms and Dynamics
- Non-Invasive Vital Sign Monitoring
- Context-Aware Activity Recognition Systems
- Advanced Image Fusion Techniques
- Image and Signal Denoising Methods
- Real-time simulation and control systems
- Force Microscopy Techniques and Applications
- Hydraulic and Pneumatic Systems
- Advanced Image Processing Techniques
- Soft Robotics and Applications
- Air Quality and Health Impacts
- Advanced oxidation water treatment
- Economic and Technological Innovation
- Innovation Diffusion and Forecasting
- Graphene research and applications
- Mechanics and Biomechanics Studies
- Refrigeration and Air Conditioning Technologies
- Fiber-reinforced polymer composites
- Polymer crystallization and properties
- Water Treatment and Disinfection
- Robotic Locomotion and Control
- Web Data Mining and Analysis
- Nanopore and Nanochannel Transport Studies
- Magnesium Alloys: Properties and Applications
Harbin Institute of Technology
2012-2024
Heilongjiang Institute of Technology
2014
Abstract Due to strong learning ability, convolutional neural networks (CNNs) have been developed in image denoising. However, operations may change original distributions of noise corrupted images, which increase training difficulty Using relations surrounding pixels can effectively resolve this problem. Inspired by that, we propose a robust deformed denoising CNN (RDDCNN) paper. The proposed RDDCNN contains three blocks: deformable block (DB), an enhanced (EB) and residual (RB). DB extract...
Photolysis via vacuum ultraviolet (VUV) irradiation is a robust technology capable of inactivating pathogens and degrading micropollutants, therefore, its application has recently attracted great interest. However, VUV water may yield nitrite (NO2–, regulated carcinogenic contaminant) hydrogen peroxide (H2O2, compound linked to aging, inflammation, cancer), thus motivating us better understand risks. By applying novel H2O2 detection method insensitive coexisting compounds, this study clearly...
This letter presents a novel atomic force microscopy (AFM)-based nanomanufacturing method combining the tip scanning with high-precision stage movement to fabricate nanochannels ladder nanostructure at bottom by continuous fixed scan size. Different structures can be obtained according matching relation of feeding velocity and precision moving velocity. relationship was first studied in detail achieve different nanostructures bottom. Machining experiments were then performed on an aluminum...
This paper focuses on the dynamic co-evolution of technology management capability and technological in system product innovation. It is motivated by increasing attention that innovation has received recently, as it considered one most important factors can bring about competitive advantages for firms to compete with their rivals current business environment. Framing a complex adaptive focusing roles system, suitable synergistic mechanisms between make reach highest fitness under diverse...
Hydrogen peroxide (H<sub>2</sub>O<sub>2</sub>) in complex water matrix can be detected reliably by ion chromatography (IC) due to its superior separation ability.
The effects of γ‐ray radiation on PBO fibers and its composite materials have been investigated, the results showed that microstructure performance radiated improved with 30–120 kGy dose, crosslinking reactions came from polymer molecule chains micro‐fibrils. As a result, molecular weight thermal stability were increased, tendency micro‐fibrils to split‐off also decreased, hook force increased by 14.0% without tensile strength loss, nano‐compression modulus flexure fiber composites 21.0%...
Deep learning (DL)-based human activity recognition (HAR) methods have shown promise in the applications of health Internet Things (IoT) and wireless body sensor networks (BSNs). However, adapting these to new users real-world scenarios is challenging due cross-subject issue. To solve this issue, we propose ActiveSelfHAR, a framework that combines active learning's benefit sparsely acquiring informative samples with actual labels self-training's effectively utilizing unlabeled data adapt HAR...
Bolt screwing assembly task is a crucial part for robot fine manipulation. For widely applied flexible joints with six-dimensional force/torque sensor, data collected could not be seen as the contact force since it consists of internal caused by deformation manipulator joints. Force analysis and geometric to judge relative pose between end-effector environment. To this end, logistic regression method was used classify state signals. Besides, criterion that bolt has entered thread hole also...
Doubly-fed wind generator is highly sensitive to voltage variations of the grid, which poses limitations for power plants during grid integrated operation. This paper proposes a new control strategy rotor side converter doubly-fed induction (DFIG) using energy-based method based on Port-Control Hamiltonian (PCH) theory and L <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> disturbance attenuation. The proposed ride-through approach limits...
This work presents studies on application and implementation the field programmable gate array (FPGA) technology into a robot joint controller in order to improve tracking performance. The can be divided two parts designed with software hardware separately. first part is developed Nios II embedded processor which used realize velocity torque control C language. second implement current vector includes coordinate transformation, PI controller, SVPWM (Space Vector Pulse Width Modulation) some...
Bolt assembly by robots is a vital and difficult task for replacing astronauts in extra-vehicular activities (EVA), but the trajectory efficiency still needs to be improved during wrench insertion into hex hole of bolt. In this paper, policy iteration method based on reinforcement learning (RL) proposed, which problem improvement constructed as an issue RL-based objective optimization. Firstly, projection relation between raw data state-action space established, then initialization designed...
Molecular self-assembly (MSA) has been established with nano-silver/triethylenetetramine (TETA) onto M40J graphite fiber surface. MSA facilitates the surface functionalization and enabling controlled interface structure of composite. The studies have shown that on in fiber/epoxy matrix interphase plays a critical role mechanical performance Detailed analysis using enhanced Raman spectrum shows TETA being self-assembled by chemical bonds surface, additional insight into enhancing composite...
Fast readout unit (FRU) has been designed for generic control and data acquisition (DAQ) in heavy-ion nuclear experiments at the Heavy Ion Research Facility Lanzhou (HIRFL) to reduce development time, production cost, maintenance difficulties of electronics. Microtelecommunications computing architecture (MicroTCA) MicroTCA systems are compatible with FRU DAQ advanced mezzanine card (AMC) module. The can connect four front-end electronics (FEE) collection, packaging, transmission via optical...
This paper discusses the design and implementation of a Web crawler - Inar written in C++ executed on Linux. It is single-threaded base asynchronous I/O technology, which under development. describes architecture function its each component detail. For some problems that we met practice, such as URL queues design, hash algorithm proposed our solution
In this paper, a high dynamic humanoid manipulator with DSP/FPGA-FPGAs hardware structure is designed to meet the demand of high-speed response required by playing ping-pong game. The integrates several sensors in each joint unit, and controller divided into two levels. upper level mainly solves kinematics trajectory planning DSP, while lower realizes control such as vector data processing FPGA. levels communicate other multipoint low-voltage differential signaling (M-LVDS) serial bus every...
In this paper, a MIMO antenna array for 5G mobile terminal is proposed. The bandwidth was broadened by opening either L-shaped slot or square in the ground plate. covering of 3.3GHz-3.6GHz and 4.8GHz-5GHz. CST software used to simulate optimize parameters designed array. prototype manufactured tested. performance simulation shows that proposed can be terminals.
Since it is difficult to obtain a closed-form solution of inverse-kinematics for humanoid robot hand with coupled joints due its nonlinearity induced by trigonometric transcendental function, most solutions achieved in early studies were numerical ones. In this paper, an algorithm based on the power series presented, which nonlinear equation including function can be converted into algebraic equation. An approximate derived firstly; 1-D linear interpolation errors compensating, accurate then...