A5100462716- Fault Detection and Control Systems
- Privacy-Preserving Technologies in Data
- Industrial Vision Systems and Defect Detection
- Soft Robotics and Applications
- Welding Techniques and Residual Stresses
- Distributed and Parallel Computing Systems
- Mineral Processing and Grinding
- Advanced Neural Network Applications
- Anomaly Detection Techniques and Applications
- Parallel Computing and Optimization Techniques
- Control and Stability of Dynamical Systems
- Robot Manipulation and Learning
- Control Systems and Identification
- Energy Harvesting in Wireless Networks
- Robotic Mechanisms and Dynamics
- Innovative Energy Harvesting Technologies
- Advanced Control Systems Optimization
- Space Satellite Systems and Control
- Stability and Control of Uncertain Systems
- Inertial Sensor and Navigation
- Target Tracking and Data Fusion in Sensor Networks
- Network Security and Intrusion Detection
- Advanced Statistical Process Monitoring
- Adaptive Control of Nonlinear Systems
- Medical Imaging Techniques and Applications
Key Laboratory of Guangdong Province
2021-2025
Sun Yat-sen University
2020-2025
Harbin Institute of Technology
2005-2024
Nanjing University of Posts and Telecommunications
2022-2024
Nanyang Technological University
2022-2024
Eye & ENT Hospital of Fudan University
2021-2024
Inner Mongolia Medical University
2024
Shenzhen University
2022-2024
Queensland University of Technology
2024
State Key Laboratory of Medical Neurobiology
2024
BackgroundMedical artificial intelligence (AI) has entered the clinical implementation phase, although real-world performance of deep-learning systems (DLSs) for screening fundus disease remains unsatisfactory. Our study aimed to train a clinically applicable DLS diseases using data derived from real world, and externally test model photographs collected prospectively settings in which would most likely be adopted.MethodsIn this national evidence study, we trained DLS, Comprehensive AI...
We have developed a genetic algorithm (GA) approach to the problem of task scheduling for multiprocessor systems. Our requires minimal specific information and no operators or repair mechanisms. Key features our system include flexible, adaptive representation an incremental fitness function. Comparison with traditional methods indicates that GA is competitive in terms solution quality if it has sufficient resources perform its search. Studies nonstationary environment show able...
In this paper, the fuzzy adaptive control problem for a class of switched stochastic nonlinear systems in pure feedback form with output constraint is addressed. By proposing mapping, constrained system transformed into an unconstrained one, equivalent objective. All signals closed-loop are proved to be semiglobally uniformly ultimately bounded. Meanwhile, satisfied and tracking error converges arbitrarily small neighborhood zero. Finally, applicability proposed controller verified by...
The issue of quality-related fault detection has attracted much attention in recent years. Partial least squares (PLS) is considered as an efficient tool for predicting and monitoring. However, due to the fact that PLS performs oblique projection input space, it not suitable detection. On other hand, a static method which cannot be used dynamic systems. In this paper, approach developed by using structure auto-regressive moving average exogenous (ARMAX) time-series model. Furthermore,...
Ā = the local vertical and horizontal reference frame B body-xed J inertia matrix of spacecraft ω angular velocity vector (rad/s) q, q 4 quaternion τ control torque input (Nm) ideal model q4 θ adaptive parameter δ
This work presents a novel bird-shaped broadband piezoelectric energy harvester based on two-DOF crossed beam for low-frequency environmental vibrations. The features cantilever mounted double-hinged beam, whose rotating motions effectively diminish its natural frequencies. Numerical simulation the finite element method is conducted to analyze modal shapes and harmonic response of proposed harvester. Prototypes are fabricated experiments carried out by testing system, results indicate good...
In surgery, the robot can automatically adjust endoscope to ensure stability of surgical image and reduce burden on doctor. It is becoming a research hotspot current assistance. The identification localization endoscopic instruments tips key achieve efficient "doctor–robot" collaboration. However, traditional methods have problems poor real-time performance insufficient information expression. This article proposes an autonomous recognition multiple instrument based arrow object bounding box...
Adeno-associated virus (AAV)-mediated gene therapy is widely applied to treat numerous hereditary diseases in animal models and humans. The specific expression of AAV-delivered transgenes driven by cell type-specific promoters should further increase the safety therapy. However, current methods for screening are labor-intensive time-consuming. Herein, we designed a "multiple vectors one AAV" strategy promoter construction vivo. Through this strategy, truncated native Myo15 hair cells (HCs)...
The steel tube arch rib in a large-span concrete-filled bridge has large span and diameter, which also leads to larger weld seam scale.Large-scale welding seams will inevitably cause more obvious residual stress (WRS).For the purpose of studying influence WRS from large-scale on mechanical properties during splicing, test research numerical simulation analysis splicing based Guangxi Pingnan Third Bridge, is world's largest bridge, were conducted this paper, distribution pattern at joint was...
ABSTRACT The state estimation of non‐cooperative spacecrafts is a crucial prerequisite for on‐orbit services. Aiming at the challenges in fusion‐based scheme with monocular vision and sparse point cloud, an optimization‐based method geometric enhancement motion proposed this paper. First, novel idea shape representation using simple features, real‐time segmentation framework established. Differing from models, it can guarantee both complete high inference speed. Second, given assumption...
The existing deep learning-based fault prognostic methods require massive labeled condition monitoring (CM) data to train a well-generalized model. However, acquiring CM for real-case machines is infeasible due time, economic costs, and safety concerns. Fortunately, we can handily obtain from relevant but different such as accelerated degradation experiments in laboratories, which contain partially shared prognosis knowledge correlated machines. Accordingly, bridge this practical gap, novel...
The use of space robots (SRs) for on-orbit services (OOSs) has been a hot research topic in recent years. However, the unstructured environment (i.e.: confined spaces, multiple obstacles, and strong radiation interference) greatly restricted application SRs. coupled active-passive multilink cable-driven robot (CAP-MCDSR) characteristics slim body, flexible movement, electromechanical separation, which is very suitable extreme environments. dynamic stiffness modeling CAP-MCDSRs challenging,...
In this article, an event-triggered active disturbance rejection control (ET-ADRC) method is designed for the battery-supercapacitor hybrid energy storage system (HESS) in electric vehicles (EVs). The proposed combines advantages of ADRC and ET mechanism. It inherits fast response from ADRC-based module, which has inner proportional–integral (PI) current loop outer voltage loop. To improve computational efficiency, module adopted by utilizing state information at last triggering sampling...
Federated learning (FL) enables multiple data owners to build machine models collaboratively without exposing their private local data. In order for FL achieve widespread adoption, it is important balance the need performance, privacy-preservation and interpretability, especially in mission critical applications such as finance healthcare. Thus, interpretable federated (IFL) has become an emerging topic of research attracting significant interest from academia industry alike. Its...
When network induced delays are considered in the event-triggered control literature, they typically from plant to controller and a tight bound on admissible is usually imposed based analysis of inter-event time. In [19], dynamic output feedback scheme introduced for stabilization input feed-forward passive (IF-OFP) networked systems (NCSs), which more general case compared with (OFP) studied [18]. Based results shown we propose set-up NCSs allows us consider both plant. We show that...
Federated learning (FL) provides a privacy-preserving approach for collaborative training of machine models. Given the potential data heterogeneity, it is crucial to select appropriate collaborators each FL participant (FL-PT) based on complementarity. Recent studies have addressed this challenge. Similarly, imperative consider inter-individual relationships among FL-PTs where some engage in competition. Although literature has acknowledged significance scenario, practical methods...
Multilink cable-driven hyper-redundant manipulators (MCDHMs) have a slender body and flexible movement characteristics, which are very suitable for complex unstructured environments. Since MCDHMs multiple coupling between active cables, linkage joints, the end-effector, mathematical modeling of segmented MCDHM becomes more complicated, uncertainty its model makes motion accuracy end-effector become low. Traditional calibration methods that rely solely on eye-in-hand or eye-to-hand method to...
Accurate and fast parameter estimation is essential for efficient operation control of permanent magnet synchronous machine (PMSM). In this paper, a novel multi-parameter decoupling method based on multi-state measurement proposed to analyze estimate the important parameters including winding resistance, magnetic flux linkage, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$d-q$</tex-math></inline-formula>...