检索结果-内蒙古大学图书馆

5th National and the 1st International Conference on Applied Research in electrical engineering, AREE 2025

作者： Azhdari, Yasin Farhang, Mahmoud Shiraz University Dept. of Electrical and Computer Engineering Shiraz Iran

ISBN: (纸本)9798350377170

Low-power and low-complexity Direction-of-Arrival (DOA) estimation is a critical challenge in various signal processing applications. This paper presents a data-driven approach using a neural network framework designed for single-snapshot, one-bit quantized measurements, enabling significant hardware simplification. We explore a Fully Connected (FC) Neural Network architecture (which is identical to a 1D convolutional architecture in the single-snapshot case) trained end-to-end on raw quantized data. The performance of this learned approach is rigorously compared against the well-established MUltiple SIgnal Classification (MUSIC) and Complex-valued Binary Iterative Hard Thresholding (CBIHT) algorithms. Our findings reveal that the proposed neural network achieves superior estimation performance compared to these, especially under low SNR values and in the presence of realistic array imperfections. We observe significant performance gains even without explicitly training the network on data with such imperfections, highlighting the generalization capabilities of the proposed method. © 2025 IEEE.

关键词： Signal to noise ratio

来源：评论

学校读者我要写书评

暂无评论

Learning-Based Ultra-Wideband Indoor NLOS Identification

Learning-Based Ultra-Wideband Indoor NLOS Identification

引用

2024 IEEE Pacific Rim Conference on Communications, computers and Signal Processing, PACRIM 2024

作者： Hu, Baojun Li, Xin Dong, Xiaodai University of Victoria Dept. Electrical and Computer Engineering Canada

ISBN: (纸本)9798350362312

This paper focuses on improving the identification accuracy of the line-of-sight (LOS) and non-line-of-sight (NLOS) conditions by channel impulse response (CIR) using machine learning methods. We propose a new definition for soft and hard NLOS based on ranging error percentages. Our research shows that accurately differentiating these two NLOS scenarios significantly enhances the LOS/NLOS identification accuracy. Experiment results demonstrate notable improvements in LOS/NLOS classification accuracy, with the accuracy of LOS detection improving from 72.33% to 93.84% using a convolutional neural network (CNN). In differentiating between hard NLOS and LOS, our accuracy reaches up to 99.82%. © 2024 IEEE.

关键词： Ultra-wideband (UWB)

来源：评论

学校读者我要写书评

暂无评论

A Study on the Selection of Torque Density Based on the Motor Output of MGM With Outer Rotor Motor and the Gear Ratio of Magnetic Gear

引用

Transactions of the Korean Institute of electrical Engineers 2025年第4期74卷 568-574页

作者： Byeon, Beom-Seok Park, Eui-Jong Kim, Yong-Jae Dept. of Electrical Engineering Chosun University Korea Republic of Dept. of Engineering for Smart Mobility Convergence Systems Chosun University Korea Republic of

In the magnetically geared machine with an inner stator(GIS) type, a magnetic geared machine, it has been observed that a significant discrepancy arises between the motor's maximum output torque and the gear's maximum input torque when the gear ratio of the magnetic gear is low. This issue persists even when the final output through the magnetic gear remains the same, due to the extensive use of permanent magnets and iron in the magnetic gear. Therefore, this paper aims to explore a method for selecting the torque density of GIS-type magnetic geared machines(MGM) by intentionally reducing the torque density of the magnetic gear. © The Korean Institute of electrical Engineers.

关键词： Permanent magnets

来源：评论

学校读者我要写书评

暂无评论

Evaluating Visual Transformers for Wafer Defect Detection: A Feasibility Study

Evaluating Visual Transformers for Wafer Defect Detection: A...

引用

2025 IEEE SoutheastCon, SoutheastCon 2025

作者： Lafrance, Samantha Bernadin, Shonda FAMU-FSU College of Engineering Dept. of Electrical and Computer Engineering Tallahassee United States

ISBN: (纸本)9798331504847

Wafer defect detection is a critical aspect of semi-conductor manufacturing, where even minor defects can lead to significant performance issues or device failures. Traditional methods, such as rule-based approaches and Convolutional Neural Networks (CNNs), have achieved moderate success but face limitations in scalability, adaptability, and the ability to model complex spatial patterns in high-resolution wafer imagery. These limitations make it critical to investigate advanced architectures that can accurately capture complex spatial relationships. Vision Transformers (ViTs) have emerged as a promising alternative within computer vision, offering global attention techniques that are particularly effective in recognizing long-range dependencies. This study evaluates the feasibility of ViTs for wafer defect detection through a literature-based empirical validation approach, synthesizing findings from prior research on ViT-based classification, defect localization, and industrial feasibility. Results indicate that ViTs offer superior accuracy and interpretability compared to CNN s, particularly in detecting rare and mixed-type defects. However, challenges remain in terms of computational cost and real-time deployment in high-throughput manufacturing settings. The findings of this study underscore the potential of ViTs as a transformative technology for semiconductor defect inspection while highlighting key areas for further development. © 2025 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Adversarial Machine Learning Attack on Machine Learning-based Controller for Solar Inverters

Adversarial Machine Learning Attack on Machine Learning-base...

引用

2024 IEEE Energy Conversion Congress and Exposition, ECCE 2024

作者： Massa, Joaquin Kim, Kyoung-Tak Kim, Taesic Ahn, BoHyun Venayagamoorthy, Kumar Zeng, Jianwu Texas A&m University-Kingsville Dept. of Electrical Engineering and Computer Science Kingsville United States Soongsil University Dept. of Electrical Engineering Seoul Korea Republic of University of Missour-Columbia Dept. of Electrical Engineering and Computer Science Columbia United States Clemson University Dept. of Electrical and Computer Engineering Clemson United States Minnesota State University Mankato Dept. of Electrical and Computer Engineering and Technology Mankato United States

ISBN: (纸本)9798350376067

Adversarial machine learning (ML) attacks are stealthy attacks designed to mislead the ML model results. This paper explores adversarial ML attacks that generate adversarial noisy input data in an ML-based controller in a solar inverter. Three types of ML models, long short-term memory (LSTM), gated recurrent unit (GRU)), and bidirectional-LSTM (Bi-LSTM), are designed to replace proportional-integral (PI) controller-based vector control for a solar inverter and two white-box adversarial ML attacks (Basic Iterative Method (BIM) attack and Fast Sign-Gradient Method (FGSM)) are applied to the ML controllers. It is observed that the adversary ML attacks designed in stealthy way do not affect the PI controller, while significantly degrading performance of the ML-based controllers. Moreover, the BIM attack is more effective than FGSM and Bi-LSTM-based controller is relatively robust to the attacks compared to peer. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

MLBERT: Multi-Level Fully Binarized BERT 1

MLBERT: Multi-Level Fully Binarized BERT

引用

1st International Conference on Innovative engineering Sciences and Technological Research, ICIESTR 2024

作者： Nasab, Mahdi Mohammadi Fakhire, Mahboube Salehi, Mostafa E. Modarresi, Mehdi University of Tehran College of Engineering Electrical and Computer Engineering Dept. Tehran Iran

ISBN: (纸本)9798350348637

BERT models, which use deep learning approaches for natural language comprehension, greatly improve language processing tasks in embedded devices. Their use in such devices enables sophisticated capabilities such as real-time language translation and making interactions more natural and efficient. BERT models, with their high computational complexity and large parameter count, impose significant issues in terms of processing power and memory requirements, particularly in resource-limited devices. These problems frequently require specialized hardware or optimization approaches, such as model compression, to enable efficient deployment without sacrificing performance. Binary Neural Networks, which simplify the weights and activations to binary values, are crucial for significantly reducing computational complexity and memory usage. This makes them ideal for deployment in resource-constrained environments like embedded systems and mobile devices, enabling efficient execution of complex machine learning tasks. Multi-level binarization (MLB) in neural network models takes advantage of bit-wise operations and enhances accuracy while reducing computational and memory requirements. We propose an MLB-based framework to enhance binarized BERT models which supports MLB for embedding layers as well. The results indicate that the DynaBERT model, when compressed to 1-bit precision for both Transformer layers and the entire embedding layers, along with 3-bit MLB activations, has no accuracy loss for SST-2, loses only about 0.5% of accuracy for MRPC, and 3% accuracy drop for RTE task, which is compatible with state-of-the-art. © 2024 IEEE.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Solution to Mitigate the Disadvantages in the Design of LLC Resonant Converters for Electric Vehicle Fast Chargers with High Output Power and Wide Output Voltage Range

引用

Transactions of the Korean Institute of electrical Engineers 2025年第2期74卷 316-325页

作者： Park, Ha-Min Lee, Hyung-Woo Park, Chan-Bae Guo, Youguang Lee, Jae-Bum Dept. of Railway Vehicle & Operation System Engineering Korea National University of Transportation Korea Republic of Dept. of Electrical and Data Engineering University of Technology Sydney Australia Dept. of Railroad Electrical and Information Engineering Korea National University of Transportation Korea Republic of

As the number of electric vehicles grows, the development of fast chargers is advancing rapidly. A DC-DC converter that can deliver high output power, a fixed input voltage, and a wide output voltage range is critical for electric vehicle fast chargers. While the LLC resonant converter is commonly used in these converters, it requires a low magnetizing inductance, which increases the transformer air gap, causing additional AC losses in the windings due to the fringing effect. To mitigate these issues, this study explores design solutions in LLC resonant converters. By using two transformers per module, with the primary sides connected in series and the secondary sides in parallel, the number of primary turns can be reduced, minimizing the air gap. This reduction also allows for structural changes in the windings, improving heat dissipation. Magnetic field analysis and circuit simulations were conducted to validate the proposed DC-DC converter design. Copyright © The Korean Institute of electrical Engineers.

关键词： Transformer windings

来源：评论

学校读者我要写书评

暂无评论

Polarimetric Image Fusion Strategies for Improving Target Shape and Contrast in Remote Sensing Applications 21

Polarimetric Image Fusion Strategies for Improving Target Sh...

引用

Electro-Optical and Infrared Systems: Technology and Applications XXI 2024

作者： Ratliff, Bradley M. Tyo, J. Scott Taylor, Michael Prikkel, Connor Dept. Electrical & Computer Engineering Dayton University DaytonOH United States Electrical & Computer Systems Engineering Dept. Monash University Melbourne Australia

ISBN: (纸本)9781510681088

Images from passive imaging polarimeters are often displayed in terms of their intensity (s0), degree of linear polarization (DoLP), and angle of polarization (AOP). The AOP and DoLP together generally provide information about shape and surface quality of imaged objects;however, the information conveyed in these images are often difficult to interpret, especially to an analyst unfamiliar with polarization phenomenology. Furthermore, it can be challenging to train machine learning approaches with direct polarimetric information due to the limited availability of application-specific data. Augmenting s0 images with polarimetric information in the form of contrast and shape enhancement thus allows for the data to remain in the domain of traditional EO-IR intensity imagery that is more familiar for both humans and trained networks. In this work we explore several strategies for effectively fusing polarimetric information into s0 imagery to improve target shape and contrast for both human and algorithmic consumption, with a particular focus on remote sensing and target detection applications. © 2024 SPIE.

关键词： Image fusion

来源：评论

学校读者我要写书评

暂无评论

EMT-Based RDR and DWT Protection Techniques Simulation of Three-Winding Power Transformer

引用

Transactions of the Korean Institute of electrical Engineers 2025年第2期74卷 258-265页

作者： Lee, Kyung-Min Park, Chul-Won Dept. of Electrical Engineering Gangneung-Wonju National University Korea Republic of

Recently, as inverter-based VREs are connected to the power grid, fault phenomena different from those of conventional synchronous generator-based power grids are appearing. This paper was conducted as a preliminary step in the development of an improved protective relaying algorithm that can prevent malfunctions in transformer systems. First, we describe the RDR based on the second harmonic using FFT, and the transformer protection technique based on DWT. We model a 345/161/23kV, 600MVA three-winding power transformer system using real system data. After reviewing the accuracy of the result of the steady-state simulation, we analyze the transients of magnetizing inrush, winding ground fault, and winding short-circuit fault. Finally, we simulate and implement the RDR protection technique with second-harmonic restraint using FFT of PSCAD library in EMT tools, and the protection of power transformer using DWT of EMT library and verify its performance. Copyright © The Korean Institute of electrical Engineers.

关键词： Electric power transmission networks

来源：评论

学校读者我要写书评

暂无评论

Study on the Application of Starting Method of Multi-Pole High Voltage Induction Motor for Drainage Pumping Station

引用

Transactions of the Korean Institute of electrical Engineers 2025年第4期74卷 561-567页

作者： Kim, Jong-Gyeum Dept. of Electrical Engineering National Gangneung-wonju University Korea Republic of

Induction motors are widely used as a power source for rotating loads due to their simplicity of structure, robustness, long lifespan, and ease of maintenance. However, these motors can cause voltage drops in loads connected to the same line when a significant voltage drop occurs during the starting process, as high current is generated. To mitigate this issue, various starting methods are employed based on the motor’s capacity to reduce the starting current. In drainage facilities that are operated temporarily during the rainy season, low-head pumping facilities are typically used. For such drainage pumping stations, which require the instantaneous discharge of large volumes of water, high-capacity multi-pole induction motors are often employed. When the motor capacity is large, reactor starting is frequently chosen over direct starting. Additionally, it is essential to consider the duration of the voltage drop when selecting the starting method. In drainage facilities, the facility capacity may be large due to temporary operations, but the operating system is relatively simple. This study compared and analyzed changes in voltage and the time required to reach the rated speed when pumping equipment, considering the moment of inertia, was started directly using a multi-pole induction motor and when started with a reactor. The analysis confirmed a difference in response time, and the voltage drop was found to be within the allowable range. © The Korean Institute of electrical Engineers.

关键词： Pumps

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：