This paper addresses the need for developing an inclusive social cyber vulnerability (iSCV) metric to assess the unique risks faced by underrepresented groups in cyberspace. Current metrics often fail to capture the v...
详细信息
A fault diagnosis method based on improved Pearson correlation algorithm and Convolutional Neural Network (CNN) is proposed to address the issue of high diagnostic speed and high proportion of semiconductor power tran...
详细信息
The benefits of new technology are becoming increasingly apparent to organisations as digital transformation continues. However, as technology becomes more widely used, cybersecurity threats and attacks are also becom...
详细信息
This paper introduces a fault-tolerant control strategy for delayed PDE-ODE coupled systems. By using a Lyapunov functional (LF), a fault-tolerant control design is developed by space-dependent linear matrix inequalit...
详细信息
Under spatially point measurements (SPMs), this paper introduces a fault-tolerant event-triggered sampled-data (FETSD) control for the output synchronization of reaction-diffusion neural networks (RDNNs). Considering ...
详细信息
A robot-assisted force control system for stable ultrasound imaging has been developed for abdomen intervention. The system aims to integrate 6-DoF robot arm, 6-axis force/torque sensor, and ultrasound probe, featurin...
详细信息
ISBN:
(纸本)9798350355376;9798350355369
A robot-assisted force control system for stable ultrasound imaging has been developed for abdomen intervention. The system aims to integrate 6-DoF robot arm, 6-axis force/torque sensor, and ultrasound probe, featuring automatic stabilization during intervention and real-time compensation for respiratory disturbance. Following the procedural workflow, the system has two operational modes. The first, rooted in admittance control, swiftly positions the robotic-held ultrasound for efficient registration. The second mode, employing the proposed adaptive control, ensures stable contact despite respiratory motion influences, enhancing procedural resilience and effectiveness. The adaptive controller predicts and eliminates disturbances more effectively than the baseline admittance controller, thanks to its utilization of the pseudo-periodic nature of respiration. The efficacy of the adaptive controller has been verified through experiments with a sponge abdomen phantom, demonstrating significant improvements in force regulation.
Deep learning is currently the mainstream method for ceramic defect detection, and it requires a large number of defect samples to train the network. However, collecting these defect samples is very time-consuming and...
详细信息
ISBN:
(纸本)9798350311259
Deep learning is currently the mainstream method for ceramic defect detection, and it requires a large number of defect samples to train the network. However, collecting these defect samples is very time-consuming and deep learning suffers from few-shot learning problems. In this study, a StyleGAN3-based data augmentation method for ceramic defect detection was proposed which can generate ceramic defect samples and thus reduce the data collection work. Experiments show that our method uses less training time, has a more stable training process, and can improve the accuracy of the detection network.
作者:
Poleshchenko, D.A.Kovrizhnykh, O.A.
National University of Science and Technology MISIS Faculty of Automation and Information Technology Stary Oskol Russia
The article describes the problem of operational planning of steelmaking production. The analysis of the production enterprise from the point of view of system approach is considered and the situational approach is ap...
详细信息
3D printing technology has rapidly advanced, yet the variability in deployment environments poses significant challenges for monitoring systems, because the dynamic deployment environments such as changes in lighting ...
详细信息
ISBN:
(纸本)9798350358513;9798350358520
3D printing technology has rapidly advanced, yet the variability in deployment environments poses significant challenges for monitoring systems, because the dynamic deployment environments such as changes in lighting and monitoring camera status, can severely impact the effectiveness and robustness of monitoring systems. To tackle this problem in monitoring system, this paper introduces a novel approach for deep neural network (DNN) adaptation to the dynamic environments through self-supervised learning and applies it to during in-situ monitoring. Specifically, we introduce a self-supervised learning strategy that leverages the auxiliary reconstruction task during in-situ monitoring, subsequently applying self-supervised fine-tuning to classification tasks with a new imbalanced-aware classification loss. Our methodology was rigorously evaluated using a real-world dataset for 3D printing defect detection. The experimental outcomes affirm the robustness of our approach, showcasing a higher defect detection accuracy rate than baselines. This substantially mitigates the adverse effects associated with printing defects, thereby increasing the reliability and quality of 3D printing processes.
This article proposes a noncontact multi-function mechanical testing device using magnetic levitation mechanism, in which a specimen can be tested in various ways (Tension, compression, bending and torsion) while the ...
详细信息
暂无评论