A5036726122- Metal and Thin Film Mechanics
- Additive Manufacturing Materials and Processes
- Advanced machining processes and optimization
- High Entropy Alloys Studies
- Advanced materials and composites
- Diamond and Carbon-based Materials Research
- Privacy-Preserving Technologies in Data
- Maritime Navigation and Safety
- Tribology and Wear Analysis
- Metallurgy and Material Forming
- Underwater Vehicles and Communication Systems
- Power Systems and Technologies
- Aluminum Alloys Composites Properties
- Maritime Security and History
- Indoor and Outdoor Localization Technologies
- Robotic Path Planning Algorithms
- Advanced Manufacturing and Logistics Optimization
- Robotics and Sensor-Based Localization
- Soil Carbon and Nitrogen Dynamics
- Wireless Networks and Protocols
- Advanced Data and IoT Technologies
- Advanced Control and Stabilization in Aerospace Systems
- Recommender Systems and Techniques
- Titanium Alloys Microstructure and Properties
- Customer churn and segmentation
China University of Mining and Technology
2021-2025
Hangzhou Dianzi University
2015-2024
University of Alberta
2022
McMaster University
2016-2022
Shanghai University
2013-2022
Beijing Research Institute of Automation for Machinery Industry (China)
2020-2021
Henan Energy & Chemical Industry Group (China)
2019
Ministry of Education of the People's Republic of China
2015
East China Jiaotong University
2014
Chongqing University
2011
Abstract With the widespread use of Internet things(IoT) devices, it generates an enormous volume data, and is a challenge to mine IoT data value while ensuring security privacy. Federated learning decentralized approach for training located on edge such as mobile phones keeping privacy, efficiency, security. However, Non-IID (non-independent identically distributed) always greatly impacts performance global model. In this paper, we propose FedDynamic algorithm solve statistical federated...
Abstract Ship path planning plays an important role in the intelligent decision-making system which can provide navigation information for ship and coordinate with other ships via wireless networks. However, existing methods still suffer from slow low security problems. In this paper, we propose a second-order model, consists of two main steps, i.e., first-order static global dynamic local planning. Specifically, first create raster map using ArcGIS. Second, is performed on based Dyna-Sarsa(...
An intelligent maritime navigation system is expected to play an important role in the realm of Internet Vessels (IoV). As a key technology systems, vessel trajectory prediction critical IoV. Automatic identification (AIS), automated tracking system, used extensively for prediction. However, certain characteristics AIS data, such as large number anchored trajectories area, anomalous sharp turns some trajectories, and behavioral differences vessels different segments, limit accuracy. In this...
In addition to the bulk properties of a workpiece material, characteristics tribofilms formed as result material mass transfer friction surface play significant role in control. This is especially true cutting hardened materials, where it very difficult use liquid based lubricants. To better understand wear performance and formation beneficial tribofilms, this study presents an assessment uncoated mixed alumina ceramic tools (Al2O3+TiC) turning two grades steel, AISI T1 D2. Both materials...
The present study investigates the morphological evolution of carbonitrides and effect these precipitates on grain boundary pinning during pseudo-carburizing a Nb-Ti-Al microalloyed steel. result indicated that three kinds complex with different morphologies containing Nb, Ti, Al respectively were observed in samples austenitized at temperatures times. NbC TiN played an important role boundaries suppressing growth austenite grains, relying high thermal stability small size precipitates....
Abstract A new method is reported for monitoring respiratory activity using photoplethysmography (PPG) without the need of prior knowledge rate range. Two channels transmission mode PPG signals were collected from subject's index finger and analyzed an independent component analysis (ICA) algorithm. The was separated heart‐related pulsation in after ICA analysis. results demonstrated that could successfully extract normal simulated apnoea activities. algorithm applied both those 10 young...
Identifying fishing vessel types with artificial intelligence has become a key technology in marine resource management. However, classical feature modeling lacks the ability to express time series features, and extraction is insufficient. Hence, this work focuses on identification of trawlers, gillnetters, purse seiners based semantic vectors. First, we extract trajectories from massive complex historical Vessel Monitoring System data that contain large amount dirty then features...
The addition of third elements may help transform brittle D022-structured lightweight Al3Ti to a relatively ductile L12-structured (Al, M)3Ti (where M represents the elements), thus increasing ductility at expense hardness. Such transformation could benefit wear resistance alloy due improved toughness if proper balance between hardness and is achieved. In this work, D022-predominant (S-Al3Ti) an L12-predominant Fe)3Ti (S-Al67Ti25Fe8) were fabricated by arc melting. Change in resistance,...
Machining quality and productivity of superalloys are limited due to their poor machinability, fewer studies have focused on the cutting iron-based superalloys. In this study, performance coated uncoated carbide tools in high-speed dry turning superalloy GH2132 was investigated by performing a series experiments. The experimental results indicated that temperature forces increased, while tool life decreased with increase speed from 30 100 m/min. Under relatively low speeds, flank face wear...
Federated learning is a new framework of machine learning, it trains models locally on multiple clients and then uploads local to the server for model aggregation iteratively until converges. In most cases, epochs all are set same value in federated learning. practice, usually heterogeneous, which leads inconsistent training speed clients. The faster will remain idle long time wait slower clients, prolongs time. As cost clients’ can reflect speed, be used guide dynamic setting epochs, we...
Abstract The defect detection of ceramic sanitary ware is often done manually, which inefficient and unhealthy. development deep learning technology makes a non-contact high-efficient method inspection possible. In this paper, based on VGG-16 network, we proposed an optimized for detection. We carried out pre-processing to denoise enhance the original image data, improved activation function MReLU, used transfer train model. results test defection data set showed that accuracy can reach...
Hard dry machining of AISI T1 and D2 steels at comparable hardness (around 59 HRC) were performed 50 m/min, 80 m/min 100 cutting speeds separately with uncoated alumina ceramic tool inserts (Al 2 O 3 + TiC), in order to study the effect conditions on generation tribofilms. Comprehensive assessment different was made by optical microscope, scanning electron microscope X-ray photoelectron spectroscopy. It shown that higher increased formation thermal protective and/or lubricating In steel,...
Considering the influence of uncertain map information on multirobot SLAM problem, a FastSLAM algorithm based landmark consistency correction is proposed. Firstly, electromagnetism-like mechanism introduced to resampling procedure in single-robot FastSLAM, where we assume that each sampling particle looked at as charged electron and attraction-repulsion electromagnetism field used simulate interactive force between particles improve distribution particles. Secondly, when multiple robots...
Producing light-weight refractory high entropy alloy (RHEA) coatings through laser cladding faces challenges like porosity and dilution, affecting their mechanical properties. Due to the inherent variability in processing parameters, accurately predicting coating morphology properties is crucial for achieving a desirable balance. This study investigates influence of power (LP), scanning speed (SS), powder feeding (PF) on porosity, micro-hardness. To this purpose, machine learning (ML)...