检索结果-内蒙古大学图书馆

Traffic Density Estimation by Distributed Proxy Model Learning for Internet-of-Vehicle

IEEE Internet of Things Journal 2025年第12期12卷 19972-19980页

作者： Li, Qilei Cheng, Jing-An Gao, Mingliang Chen, Jinyong Jeon, Gwanggil Shandong University of Technology School of Electrical and Electronic Engineering Zibo255000 China Queen Mary University of London School of Electronic Engineering and Computer Science LondonE1 4NS United Kingdom Incheon National University Department of Embedded Systems Engineering Incheon22012 Korea Republic of

Autonomous driving has been significantly advanced in todays society, which revolutionized daily routines and facilitated the development of the Internet of Vehicles (IoV). A crucial aspect of this system is understanding traffic density to enable intelligent traffic management. With the rapid improvement in deep neural networks (DNNs), the accuracy of density estimation has markedly improved. However, there are two main issues that remain unsolved. Firstly, current DNN-based models are excessively heavy, characterized by an overwhelming number of training parameters (millions or even billions) and substantial computational complexity, indicated by a high number of FLOPs. These requirements for storage and computation severely limit the practical application of these models, especially on edge devices with limited capacity and computational power. Secondly, despite the superior performance of DNN models, their effectiveness largely depends on the availability of large-scale data for training. Growing privacy concerns have made individuals increasingly hesitant to allow their data to be publicly used for model training, particularly in vehicle-related applications that might reveal personal movements, which leads to data isolation issues. In this paper, we address these two problems at once with a systematic framework. Specifically, we introduce the Proxy Model Distributed Learning (PMDL) model for traffic density estimation. PMDL model is composed of two main components. First, we introduce a proxy model learning strategy that transfers fine-grained knowledge from a larger master model to a lightweight proxy model, i.e., a proxy model. Second, we design a distributed learning strategy that trains multiple proxy models with privacy-aware local data and seamlessly aggregates these models via a global parameter server. This ensures privacy protection while significantly improving estimation performance compared to training models with limited, isolated data. We tested

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

AlN/GaN/AlN HEMTs on Bulk AlN Substrates with High Drain Current Density > 2.8 A/mm and Average Breakdown Field > 2 MV/cm 82

AlN/GaN/AlN HEMTs on Bulk AlN Substrates with High Drain Cur...

引用

82nd Device Research Conference, DRC 2024

作者： Kim, Eungkyun Chen, Yu-Hsin Encomendero, Jimy Jena, Debdeep Xing, Huili Grace Cornell University School of Electrical and Computer Engineering IthacaNY14853 United States Cornell University Department of Materials Science and Engineering IthacaNY14853 United States

ISBN: (纸本)9798350373738

Wide energy bandgap and high saturation velocity of GaN have played a pivotal role in increasing the output power density (Pout) of a radio frequency (RF) transistor, allowing for an increased radar and communication system range. A further improvement in output power densities of GaN HEMTs is currently thermally limited due to the waste heat generated during the device operation in the channel layer. Therefore, to realize reliable and sustained operation at high Pout , thermal management in GaN HEMTs must be improved. We present AlN/GaN/AlN HEMTs, where an AlN buffer layer is homoepitaxially grown on bulk AlN substrates. In this heterostructure, AlN simultaneously offers an ultra-wide energy bandgap (~6.1 eV) and high thermal conductivity (340 W/m·K) [1];the thermal advantage of an AlN buffer layer in HEMTs grown on SiC substrates has been recently confirmed by Soctera [2]. AlN homoepitaxy employed in our heterostructure further improves the thermal performance by potentially eliminating the thermal boundary resistance at the growth interface [3]. To this end, we report the on-current and breakdown behavior of HEMTs grown on Al-polar bulk AlN for the first time after the dramatically improved AlN homoepitaxy by MBE [4]. © 2024 IEEE.

关键词： Aluminum nitride

来源：评论

学校读者我要写书评

暂无评论

Optimizing Deep Learning Inference to Detect PCB Soldering Defects 4

Optimizing Deep Learning Inference to Detect PCB Soldering D...

引用

4th IEEE International Conference on Artificial Intelligence in engineering and Technology, IICAIET 2022

作者： Tham, Mau-Luen Chong, Boon Yoong Tan, Yu Han Wong, Yen Khai Chean, Swee Ling Tan, Wei Kun Department of Electrical and Electronic Engineering Malaysia Department of Internet Engineering and Computer Science Malaysia

ISBN: (数字)9781665468374

ISBN: (纸本)9781665468374

Rapid and accurate defect detection in printed circuit board (PCB) manufacturing plays a vital role in the quality control (QC) of consumer electronic products. Automated visual inspection is gaining increased popularity owing to low production cost and high portability. Prior works rely on a combination of image subtraction and processing techniques whereas the recent works utilize the generalization capability of deep learning. The former operates poorly in a mismatch condition of image pairs whereas the latter consumes higher computation time and processing power. In this paper, we propose a YOLOv4-based PCB detection framework which detects six types of defects. To attain production-ready capability, we optimize the models using the Intel OpenVINO toolkit. Experiment results reveal that the optimized models yield good accuracy for defect detection with significant execution speed on low-power embedded platform. © 2022 IEEE.

关键词： Printed circuit boards

来源：评论

学校读者我要写书评

暂无评论

Predictive modeling of MXene-based solar absorbers using a deep neural network

引用

Journal of the Optical Society of America B: Optical Physics 2025年第4期42卷 763-772页

作者： Waseer, Waleed Iqbal Baqir, Muhammad Abuzar Saqlain, Muhammad Mughal, Muhammad Junaid Khan, Shujaat School of Physics State Key Laboratory of Electronic Thin Films and Integrated Devices University of Electronic Science and Technology of China Chengdu China Department of Electrical Engineering COMSATS University Islamabad Sahiwal57000 Pakistan Department of Electrical and Computer Engineering COMSATS University Islamabad Islamabad45550 Pakistan Department of Computer Engineering College of Computing and Mathematics King Fahd University of Petroleum & Minerals Dhahran31261 Saudi Arabia SDAIA-KFUPM Joint Research Center for Artificial Intelligence King Fahd University of Petroleum & Minerals Dhahran31261 Saudi Arabia

Nanophotonic structures, such as photonic crystals, plasmonic nanostructures, and metamaterials, present transformative potential in advancing optical devices through innovative design capabilities. Among these, metamaterials—specifically metal–insulator–metal (MIM)-based absorbers—stand out for their ability to achieve high electromagnetic wave absorption across designated frequencies, proving valuable in applications ranging from solar cells to sensors. This study focuses on designing a high-efficiency MXene–SiO2–silver MIM metasurface absorber, leveraging deep learning techniques to streamline and enhance the simulation process traditionally reliant on finite-difference time-domain (FDTD) or finite-element methods (FEMs). By utilizing a deformable convolutional neural network (CNN), the proposed model predicts absorption spectra from metasurface geometries with superior accuracy and reduced computational demand while achieving dynamic spatial transformation handling. The model’s accuracy is further elucidated through Gradient-weighted Class Activation Mapping (Grad-CAM), which provides insights into the absorber’s most influential regions, supporting targeted optimization for specific applications. This deep learning approach demonstrates significant advantages over conventional methods in efficiency and adaptability, advancing the development of high-performance nanophotonic devices. © 2025 Optica Publishing Group. All rights, including for text and data mining (TDM), Artificial Intelligence (AI) training, and similar technologies, are reserved.

关键词： Time domain analysis

来源：评论

学校读者我要写书评

暂无评论

Early Detection of Fire in EV Battery Using Machine Learning Approach 2

Early Detection of Fire in EV Battery Using Machine Learning...

引用

2nd Edition IEEE International Conference on Metrology for eXtended Reality, Artificial Intelligence and Neural engineering, MetroXRAINE 2023

作者： Kaur, Ishpreet Singh, Mukesh Kasana, Singara Singh Thapar Institute of Engineering and Technology Computer Science and Engineering Department Patiala India Thapar Institute of Engineering and Technology Electrical and Instrumentation Engineering Department Patiala India

ISBN: (纸本)9798350300802

In recent years, Electric Vehicles (EVs) have revolutionized the automobile industry. Nowadays, EVs are the preferred means of transportation for people because they are easy to drive, convenient, and make less noise. EVs use Lithium-Ion Batteries (LIB) to power the vehicle. However, in past years, there have been multiple instances of EVs catching fire and becoming one of the most urgent problems in EVs. The leading causes of such mishaps include faulty Battery Management System (BMS), no proper venting mechanism for overheated cells to release heat, poor quality cells, and short-circuiting. Many researchers proposed fire handling systems that consider sensor values, such as temperature, voltage, etc. while some focus on battery health management. However, there is a need for a detection mechanism to detect fire in advance to save people's life and property from damage. Some researchers have proposed fire detection strategies considering only a few sensor readings that cannot detect the extent of a fire early and inform users about the situation. Therefore, in the proposed work, an approach is designed considering both environmental and battery data using multi-sensors for developing an intelligent early fire detection system. The proposed approach first performs exploratory data analysis on real dataset to get an insight into data for fault diagnosis, leading to early thermal runaway. Further, Machine Learning (ML) and ensemble methods are used to check the potentiality of the fire or detect fire in advance. Finally, the system is evaluated concerning accuracy, recall, precision, and Fl-Score. © 2023 IEEE.

关键词： Battery Management System Electric Vehicles Fire Detection Lithium-Ion Battery Machine Learning

来源：评论

学校读者我要写书评

暂无评论

Internet of Medical Things (IoMT) for Premature Estimate of Epileptic Seizures 16

Internet of Medical Things (IoMT) for Premature Estimate of ...

引用

16th International Conference on Developments in eSystems engineering, DeSE 2023

作者： Alrawi, Ali Amer Ahmed Mashhadany, Yousif Al Najeeb, Mushtaq Hadi, Sura A. Algburi, Sameer Graovac, Stevica University of Anbar Department of Electrical Engineering Anbar Iraq University Science Malaysia School of Computer Science Malaysia Al-Kitab University College of Engineering Kirkuk Iraq University of Belgrade Faculty of Electrical Engineering Belgrade Serbia Faculty of Electrical Engineering Montenegro

ISBN: (纸本)9798350381344

This paper proposes an Internet of Medical Things (IoMT) Seizure Detection Algorithm that uses smartphone acceleration sensors to detect early seizures. The propose algorithm used based on MATLAB Mobile, which seamlessly accesses MATLAB capabilities and uses MATLAB Drive for cloud-based data storage and analysis. Time-domain and frequency-domain analysis turn acceleration impulses into meaningful feature vectors. Then, machine learning classifiers like SVMs or deep learning models are used to train the system on a seizure and non-seizure dataset. The algorithm is calibrated to distinguish s suraabdlattif@*** eizure occurrences from regular activity. The IoMT technique uses real-time streaming to send sensor data to the cloud-based MATLAB Drive for continuous monitoring. The MATLAB Mobile app allows smartphone users to remotely monitor seizure activity using the seizure detection algorithm. The proposed IoMT Seizure Detection Algorithm is tested utilizing a broad dataset of seizure episodes and physical activities. Accuracy, sensitivity, specificity, and the ROC curve are evaluated. The algorithm's robustness and effectiveness in detecting seizure occurrences highlight its potential as a noninvasive, cost-effective early seizure detection system. MATLAB Mobile and MATLAB Drive integration shows that IoMT-based healthcare solutions implemented in real-world circumstances, improving patient care and medical insights. The auto-correlation between the patient's motion and the [] data helped clinicians decide if the patient's motion was typical or not. Up to 85% consistency was found while using our proposed method for detecting epilepsy. © 2023 IEEE.

关键词： Smartphones

来源：评论

学校读者我要写书评

暂无评论

A Detailed Comparative Study of AI-Based Intrusion Detection System for Smart Cities

A Detailed Comparative Study of AI-Based Intrusion Detection...

引用

2024 International Conference on Communication, computer sciences and engineering, IC3SE 2024

作者： Rakha, Muhammad Allah Akbar, Abdullah Chhabra, Gunjan Kaushik, Keshav Arshi, Oroos Khan, Inam Ullah Fast National University Of Computer And Emerging Sciences Department Of Computer Science Peshawar Pakistan National University Of Computer And Emerging Sciences Department Of Electrical Engineering Peshawar Pakistan Graphic Era Hill University Department Of Cse Dehradun India Graphic Era Deemed To Be University Uttarakhand Dehradun India School Of Computer Science University Of Petroleum And Energy Studies Dehradun India Isra University Department Of Electronic Engineering Islamabad Campus Pakistan

ISBN: (纸本)9798350366846

The proliferation of Internet of Things (IoT) devices poses potential challenges in the fast-developing field of smart cities, especially in cybersecurity. This work is an attempt to present an extensive comparative analysis of AI-based intrusion detection systems (IDS) designed for smart cities. The necessity of robust intrusion detection systems is highlighted early on in discussing the security threats associated with IoT and its importance in enhancing urban life. The study looks at current datasets that are frequently utilized to train IDS models, emphasizing both their shortcomings and the growing need for more diverse and representative datasets. This work explores the use of deep learning (DL) and machine learning (ML) approaches for intrusion detection in smart city environments, it looks at different classifiers and how well they identify cyber threats. This paper presents simulation methods and experimental results for demonstrating the accuracy, precision, recall, and F1-score performance of many classifiers using the complex datasets NF-BoT-IoT and IoTID20. This paper also touches upon the significance of continuing research in utilizing cutting-edge algorithms and novel datasets to improve smart cities' cybersecurity posture and, eventually, aid in creating more complex intrusion detection systems that can successfully handle the ever-changing threat landscape. It also emphasizes the necessity of ongoing innovation in this field and the potential for deep learning algorithms to improve cyberattack detection skills. © 2024 IEEE.

关键词： Smart city

来源：评论

学校读者我要写书评

暂无评论

Shape Organization for Enhanced Design Optimization in Generative Systems Using Clustering Technology 6

Shape Organization for Enhanced Design Optimization in Gener...

引用

6th International Conference on Recent Trends in Advance Computing, ICRTAC 2023

作者： Krishna, Raji Suganya, R. Yaganti, Mounika Sai Mohamed Buhari, Seyed Vellore Institute of Technology School of Electrical Engineering Chennai Campus Chennai India Vellore Institute of Technology School of Computer Science Engineering Chennai Campus Chennai India University Brunei Darussalam Department of Computer Science Gadong Brunei

ISBN: (纸本)9798350394702

The current research in design space articulation tries to solve the problem that comes up because computer generative systems are becoming more complex, and there are more design options. This study makes a contribution to the existing body of research by introducing a technique for form clustering that utilizes K-medoids, a Machine Learning approach. The suggested method uses a clustering method to group design shapes that are similar and then pick a representative body for each cluster in a 2D-Grid Based representation. The efficacy of the process has been empirically validated by applying it to a novel test case and comparing it with other clustering evaluation methods. The accuracy of the results of this evaluation is about the same as that seen in an outside study. This gives us essential information about how to judge machine learning techniques. © 2023 IEEE.

关键词： Hamming distance

来源：评论

学校读者我要写书评

暂无评论

Security, Trust, and Privacy in Cyber-physical Systems (CPS) 2

Security, Trust, and Privacy in Cyber-physical Systems (CPS)

引用

2nd International Conference on Cyber Physical Systems, Power Electronics and Electric Vehicles, ICPEEV 2024

作者： Busari, Wareez A. Bello, Abeeb A. Department of Cyber Security Science Federal University of Technology Niger State Minna Nigeria Department of Electrical and Computer Engineering Mahindra University Hyderabad India

ISBN: (纸本)9798350388664

This paper explores the convergence of physical and computational components in Cyber-Physical Systems (CPS). Security, trust, and privacy are paramount for reliable and resilient operation of these interconnected systems across diverse sectors like healthcare and smart cities. The paper analyses the complex interplay of these factors, identifying practical challenges and proposing solutions to enhance system robustness. By analysing trends, research gaps, and offering actionable recommendations, the study aims to empower stakeholders in fortifying CPS against evolving cyber threats. It emphasizes the need for comprehensive strategies to mitigate risks, protect sensitive information, and ensure data integrity and system reliability. Through a holistic approach integrating security, trust, and privacy mechanisms, the research contributes to advancing effective security measures in CPS, ultimately fostering a culture of trust and resilience in the digital era. © 2024 IEEE.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

AbotalebNet: A Novel Deep Learning Architecture for Time Series Forecasting with Convolutional and LSTM Layers Enhanced by Multi-Head Attention 2

AbotalebNet: A Novel Deep Learning Architecture for Time Ser...

引用

2nd IEEE International Conference on Trends in Quantum Computing and Emerging Business Technologies, TQCEBT 2024

作者： Abotaleb, Mostafa Makarovskikh, Tatiana Mijwil, Maad M. Bala, Indu Hiran, Kamal Kant Doshi, Ruchi South Ural State University Dep. of System Programming Chelyabinsk454080 Russia Baghdad College of Economic Sciences University Computer Techniques Engineering Department Baghdad Iraq Lovely Professional University Electrical and Electronics Engineering Punjab India Sir Padampat Singhania University Dep. of Computer Science Udaipur India Universidad Azteca Dep. of Computer Science and Engineering Chalco Mexico

ISBN: (纸本)9798350384277

This paper introduces AbotalebNet, a novel deep learning architecture optimized for time series forecasting, with a particular focus on the complexities of COVID-19 data. AbotalebNet's architecture is mathematically represented as:\begin{equation*}\hat Y = {\text{Attention}}(\operatorname{LSTM} ({\text{CNN}}(X)))\tag{1}\end{equation*}Here, X represents the input time series data, and Ŷ is the forecasted output. AbotalebNet integrates the feature extraction capabilities of Convolutional Neural Networks (CNNs) with the sequential data processing power of Long Short-Term Memory (LSTM) networks, further enhanced by a Multi-Head Attention *** address overfitting, a key challenge in deep learning, the model incorporates regularization strategies, dropout mechanisms, and batch normalization:\begin{equation*}\mathop {\min }\limits_\Theta ( Loss (X,\hat Y;\Theta ) + \lambda R(\Theta ))\tag{2}\end{equation*}In this equation, Θ denotes the model parameters, R(Θ) represents regularization terms, and λ is the regularization coefficient. These additions aid in preventing the model from overfitting to the training data, ensuring robust performance on unseen *** evaluation on COVID-19 time series data demonstrates AbotalebNet's enhanced predictive accuracy over traditional models, solidifying its potential for advanced non-linear time series analysis. © 2024 IEEE.

关键词： Convolutional neural networks

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：