检索结果-内蒙古大学图书馆

Annual IEEE Conference on Applied Power Electronics Conference and Exposition (APEC)

作者： Ayooluwa Ajiboye Ayodhya Somiruwan Gamwari Rakesh Resalayyan Alireza Khaligh Department of Electrical and Computer Engineering Maryland Power Electronics Laboratory The Institute for Systems Research University of Maryland College Park USA

The high commutation speed of wide band-gap devices causes undesired voltage overshoots and oscillatory behavior observed across the drain-source port. The mitigation of these voltage overshoots and oscillations caused by the resonant interaction between the commutation loop parasitic inductance and the switch parasitic output capacitance requires additional design considerations. In this work, a Thevenin-based frequency domain approach for modelling the switch voltage overshoot and oscillations with respect to the commutation power loop is proposed. The switching transition and its associated slew rate is modelled using a clamped-ramp function. The two figures of merit used to quantify the effect of the circuit elements in the commutation power loop with respect to the resonant oscillations are the voltage overshoot and time to steady state. Using the clamped-ramp function and the two figures of merit, the effect of the parasitic inductance and the decoupling capacitor on the power loop is analyzed. The trends observed from the analyses are used to obtain recommendations regarding the sizing and placement of the decoupling capacitor in the power loop. The proposed Thevenin-based frequency domain model is experimentally validated by comparing the analytical results with both the LTspice simulations and the experimental waveforms for a SiC-based 400 V, 20 A double pulse test.

关键词：

来源：评论

学校读者我要写书评

暂无评论

BalanceDN: Load-Balancing Allocation of Interest for Fast Discovery in Content Centric Networks

arXiv

引用

arXiv 2024年

作者： Gunti, Murali Rojas-Cessa, Roberto Networking Research Laboratory Department of Electrical and Computer Engineering New Jersey Institute of Technology NewarkNJ07102 United States

In Named Data Networking (NDN), data is identified by unique names instead of IP addresses, and routers use the names of the content to forward Interest packets towards the producers of the requested content. However, the current content search mechanism in NDN is complex and slow. This mechanism not only creates congestion but also hinders practical deployment due to its slowness and cumbersome nature. To address this issue, we propose a methodology, called BalanceDN, that distributes content through the network such that sought content can be found quickly. BalanceDN uses a distributed allocation of resolvers as those used by the domain name system but differs in how content is distributed. Our approach avoids flooding the network with pending interest requests and also eliminates the need for blind search when the location of content is unknown. We tested our approach on a simulation platform for NDN called ndnSIM. The results show that the proposed routing scheme utilizes far fewer network resources compared to the NDN network when retrieving content. The proposed scheme accomplishes this performance gain by leveraging a load-balanced hashing mechanism to distribute and locate the name of the content on the distributed nameserver lookup service nodes. © 2024, CC BY-NC-ND.

关键词： Simulation platform

来源：评论

学校读者我要写书评

暂无评论

Review of Inference Time Prediction Approaches of DNN: Emphasis on Service robots with cloud-edge-device architecture

Review of Inference Time Prediction Approaches of DNN: Empha...

引用

2023 IEEE International Conference on Robotics and Biomimetics, ROBIO 2023

作者： Xiang, Tian Meng, Qiwei Zhang, Ji Zhang, Beibei Song, Wei Xie, Anhuan Gu, Jason Research Institute of Interdisciplinary Innovation Zhejiang Laboratory Research Center for Intelligent Robotics Hangzhou311100 China The Chinese University of Hong Kong T Stone Robotics Institute Department of Mechanical and Automation Engineering Hong Kong Research Institute of Intelligent Computing Zhejiang Laboratory Research Center for Intelligent Computing Software Hangzhou311100 China Dalhousie University Electrical and Computer Engineering Halifax Canada

ISBN: (纸本)9798350325706

In recent years, the global robot market has witnessed substantial growth, particularly in the domain of service robots. Despite their expanding presence, service robots encounter limitations when operating autonomously in unstructured environments, primarily due to their constrained computational capacities. As a solution, the fusion of cloud and edge computing resources becomes imperative to expedite task inference and enhance scenario perception capabilities. The integration of cloud-edge-device models holds significant promise in bolstering the operational efficiency of robots. This entails the dynamic partitioning of intricate robotic tasks, executed collaboratively across cloud, edge, and device resources. In this landscape, deep neural network (DNN) models play a pivotal role in facilitating a wide array of robotic tasks. The inference time for each layer of a DNN model in actual deployment, emerges as a critical determinant in the model's partitioning strategy. It also serves as an important metric influencing the model's suitability for a specific hardware platform. This article presents an overview of recent advancements in predicting inference and training time of DNN models, summarizes the related methods, and finally discusses the challenges in this field and the research that can be studied in the future. © 2023 IEEE.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Semantic Communication for Cooperative Perception Based on the Importance Map 15

Semantic Communication for Cooperative Perception Based on t...

引用

15th IEEE International Conference on Wireless Communications and Signal Processing, WCSP 2023

作者： Sheng, Yucheng Ye, Hao Liang, Le Jin, Shi University of California Department of Electrical and Computer Engineering Santa CruzCA95064 United States National Mobile Communications Research Laboratory Southeast University Nanjing210096 China Purple Mountain Laboratories Nanjing211100 China

ISBN: (纸本)9798350324662

Cooperative perception, which has a broader perception field than single-vehicle perception, has played an increasingly important role in autonomous driving to conduct 3D object detection. Through vehicle-to-vehicle (V2V) communication technology, various connected automated vehicles (CAVs) can share their sensory information (LiDAR point clouds) for cooperative perception. We employ an importance map to extract significant semantic information and propose a novel cooperative perception semantic communication scheme with intermediate fusion. Meanwhile, our proposed architecture can be extended to the challenging time-varying multi-path fading channel. To alleviate the distortion caused by the time-varying multi-path fading, we adopt orthogonal frequency-division multiplexing (OFDM) blocks combined with channel estimation and channel equalization. Simulation results demonstrate that our proposed model outperforms the traditional separate source-channel coding over various channel models. © 2023 IEEE.

关键词： Cooperative perception machine learning semantic communication time-varying multipath fading V2V communication

来源：评论

学校读者我要写书评

暂无评论

Unsupervised Anomaly Detection for Automotive CAN Bus on the Intel Loihi

Unsupervised Anomaly Detection for Automotive CAN Bus on the...

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Rashedul Islam Shahanur Alam Chris Yakopcic Nayim Rahman Simon Khan Tarek Taha Department of Electrical and Computer Engineering University of Dayton Dayton Ohio USA Air Force Research Laboratory Rome NY USA

ISBN: (数字)9798350359312

ISBN: (纸本)9798350359329

Detecting anomalies and faults swiftly and accurately is essential in many applications, such as healthcare, infrastructure, industry, and security. In most of these applications, the available power for running an anomaly detection system is very limited. This is especially the case for modern smart and electric vehicles, which have an increasing amount of electronics. These vehicles have Electronic Control Units (ECUs) that communicate with each other through the Controller Area Network (CAN) bus. The Controller Area Network, however, is not very secure, so bad actors can do severe damage, putting the lives of drivers and passengers at risk. Thus, low-power anomaly detection is essential for both security and reliability. Deep learning networks have demonstrated proficiency in anomaly identification. However, their implementation on conventional Central Processing Units (CPUs) and Graphics Processing Units (GPUs) incurs significant energy consumption. To overcome this limitation, we propose a low-power approach for placing security directly in the network hardware. An autoencoder-based real-time anomaly detector trained through unsupervised learning is seamlessly mapped onto the Intel Loihi spiking neuromorphic processor. Our proposed anomaly detector used significantly less power than a CPU or GPU and showed the best accuracy and F-1 score to detect anomalies compared to the alternative spiking approaches. The proposed anomaly detection system consumed about 20,000 and 700 times less energy than a CPU and GPU respectively. To the best of our knowledge, this is the first low-power, unsupervised anomaly detection system using the Loihi or any other neuromorphic processor. The results of this work will apply to other anomaly detection applications, as well as other spiking neuromorphic processors, resulting in a universally applicable extreme low-power anomaly detection platform.

关键词： Industries Energy consumption Neuromorphics Neural networks Graphics processing units Hardware Central Processing Unit

来源：评论

学校读者我要写书评

暂无评论

WHALE-FL: Wireless and Heterogeneity Aware Latency Efficient Federated Learning over Mobile Devices via Adaptive Subnetwork Scheduling 39

WHALE-FL: Wireless and Heterogeneity Aware Latency Efficient...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Su, Huai-An Geng, Jiaxiang Li, Liang Qin, Xiaoqi Hou, Yanzhao Wang, Hao Fu, Xin Pan, Miao Department of Electrical and Computer Engineering University of Houston United States School of Information and Communication Engineering Beijing University of Posts and Telecommunications China Frontier Research Center Peng Cheng Laboratory China Department of Electrical and Computer Engineering Stevens Institute of Technology United States

ISBN: (纸本)157735897X

As a popular distributed learning paradigm, federated learning (FL) over mobile devices fosters numerous applications, while their practical deployment is hindered by participating devices’ computing and communication heterogeneity. Some pioneering research efforts proposed to extract subnetworks from the global model, and assign as large a subnetwork as possible to the device for local training based on its full computing capacity. Although such fixed size subnetwork assignment enables FL training over heterogeneous mobile devices, it is unaware of (i) the dynamic changes of devices’ communication and computing conditions and (ii) FL training progress and its dynamic requirements of local training contributions, both of which may cause very long FL training delay. Motivated by those dynamics, in this paper, we develop a wireless and heterogeneity aware latency efficient FL (WHALE-FL) approach to accelerate FL training through adaptive subnetwork scheduling. Instead of sticking to the fixed size subnetwork, WHALE-FL introduces a novel subnetwork selection utility function to capture device and FL training dynamics, and guides the mobile device to adaptively select the subnetwork size for local training based on (a) its computing and communication capacity, (b) its dynamic computing and/or communication conditions, and (c) FL training status and its corresponding requirements for local training contributions. Our evaluation shows that, compared with peer designs, WHALE-FL effectively accelerates FL training without sacrificing learning accuracy. Copyright © 2025, Association for the Advancement of Artificia Intelligence (***). All rights reserved.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

All-silicon integrated platform for light generation, manipulation, and detection—and its application to refractive index sensors

引用

Optica 2025年第5期12卷 577-584页

作者： de Cea, Marc Kramnik, Danielius Stojanovic, Vladimir Ram, Rajeev J. Research Laboratory of Electronics Massachusetts Institute of Technology CambridgeMA02139 United States Department of Electrical Engineering and Computer Sciences UC Berkeley BerkeleyCA94720 United States

Traditional silicon photonic platforms offer powerful capabilities for light manipulation and detection but cannot realize an integrated light source. This results in systems with increased cost and complexity due to the need for external light generation and coupling. In this work we overcome this limitation and experimentally demonstrate the generation, manipulation, and detection of waveguide-coupled light in a commercial silicon photonics foundry with monolithically integrated electronics—without any modification to the fabrication process and no additional post-processing steps. We propose the use of our platform to realize fully integrated, monolithic optical refractometric sensors and theoretically show that performance comparable to commercial, state of the art bench-top systems can be achieved. Our work lays the foundation for the realization of truly monolithic, full-functionality silicon photonic electro-optical systems with increasing levels of miniaturization and cost reduction. © 2025 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement.

关键词： Silicon photonics

来源：评论

学校读者我要写书评

暂无评论

An Attention-Based Model for Clinical Time Series Prediction: Enhancing ICU Readmission Prediction

An Attention-Based Model for Clinical Time Series Prediction...

引用

International eConference on computer and Knowledge engineering (ICCKE)

作者： Hananeh Sadat Madinei Mohammad Reza Keyvanpour Seyed Vahab Shojaedini Department of Computer Engineering Data Mining Laboratory Faculty of Engineering Alzahra University Tehran Iran Department of Electrical Engineering Iranian Research Organization for Science and Technology Tehran Iran

ISBN: (数字)9798331511272

ISBN: (纸本)9798331511289

In recent years, time series prediction has become a highly interesting topic in various applied areas, including clinical time series analysis. Hospitals and other clinical healthcare systems collect Electronic Health Records. These records contain important information that needs to be mined. This information provides a solid basis for predicting hospital readmission, helping to improve healthcare, enhance patient outcomes, and optimize hospital resources. This study introduces a novel deep learning attention-based approach for predicting 30-day ICU readmissions using clinical time series information extracted from the MIMIC III dataset. This approach incorporates time-aware positional encoding into the attention mechanism to enhance predictive performance by capturing the temporal representation of data and considering the temporal order of clinical time series data. This model is compared with cutting-edge models. The significant enhancements in the model performance achieved by the proposed method highlight the effectiveness of incorporating time-aware positional encoding in capturing temporal dependencies and improving prediction outcomes. These results underscore the potential of advanced attention mechanisms in clinical time series analysis.

关键词： Deep learning Analytical models Attention mechanisms Hospitals Computational modeling Time series analysis Predictive models Encoding Data models Electronic medical records

来源：评论

学校读者我要写书评

暂无评论

A Scalable Real-Time Distributed Multiagent Decision Making Algorithm with Cost 19

A Scalable Real-Time Distributed Multiagent Decision Making ...

引用

19th IEEE Annual Consumer Communications and Networking Conference, CCNC 2022

作者： Cotae, Paul Kang, Myong Velazquez, Alexander University of the District of Columbia Department of Electrical and Computer Engineering WashingtonDC United States Center for High Assurance Computer Systems Naval Research Laboratory WashingtonDC20375 United States

ISBN: (纸本)9781665431620

We present a distributed algorithm for a multiagent system in a collaborative setting including the cost of actions. We revise the distributed Max-Plus algorithm by presenting a new Cost Distributed Max-Plus (CD Max-Plus) algorithm by including the cost of actions in the interactions of agents, named Factored Value-MCTS-CD Max-Plus algorithm. The proposed Factored Value-MCTS-CD Max-Plus algorithm is online, anytime, distributed, and scalable in number of agents and their interactions. Our contribution is an alternative solution for solving cost Dec-POMDPs competing with the state of art methods and algorithms using MCTS and Max-Plus algorithms to exploit the locality of agent interactions for planning and acting. © 2022 IEEE.

关键词： Multi agent systems

来源：评论

学校读者我要写书评

暂无评论

The spectrum analysis of RMP coils with multiple connection modes and the design of high field side coils on J-TEXT

引用

Plasma Science and Technology 2022年第12期24卷 224-231页

作者： Zhengkang REN Da Li Nengchao WANG Feiyue MAO Zhuo HUANG Song ZHOU Ruo JIA Ying HE Chengshuo SHEN Abba Alhaji BALA Bo RAO Yonghua DING the J-TEXT Team International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics State Key Laboratory of Advanced Electromagnetic Engineering and TechnologySchool of Electrical and Electronic EngineeringHuazhong University of Science and TechnologyWuhan 430074People's Republic of China College of Computer Science South-Central Minzu UniversityWuhan 430074People's Republic of China Physics Department Federal University DutseDutse 720101Nigeria

The phase difference Δξ between locked islands(2/1 and 3/1) has been found to influence the heat transport on the thermal quench during disruptions by numerical modeling [Hu Q et al *** 59,016005].To verify this experimentally,a set of resonant magnetic perturbation(RMP) coils is required to excite coupled magnetic islands with different Δξ.The spectrum analysis shows that the current RMP coils on J-TEXT can only produce sufficient 2/1 and 3/1RMP fields with a limited phase difference of Δξ∈[-75°,75°].In order to broaden the adjustable range of Δξ,a set of coils on the high field side(HFS) is proposed to generate 2/1 and 3/1 RMP fields with Δξ=180°.As a result,RMPs with adjustable Δξ∈[-180°,180°] and sufficient amplitudes could be achieved by applying the HFS coils and the low field side(LFS)*** work provides a feasible solution for flexible adjustment of the phase difference between m and m+1 RMP,which might facilitate the study of major disruptions and their control.

关键词： resonant magnetic perturbation spectrum analysis phase difference between 2/1 and 3/1 islands

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：