检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Ferenczi, Bryce Burke, Michael Drummond, Tom Department of Electrical and Computer Systems Engineering Monash University Australia Department of Computing and Information Systems Melbourne University Australia

This work introduces a novel and adaptable architecture designed for real-time occupancy forecasting that outperforms existing state-of-the-art models on the Waymo Open Motion Dataset in Soft IOU. The proposed model uses recursive latent state estimation with learned transformer-based functions to effectively update and evolve the state. This enables highly efficient real-time inference on embedded systems, as profiled on an Nvidia Xavier AGX. Our model, MotionPerceiver, achieves this by encoding a scene into a latent state that evolves in time through self-attention mechanisms. Additionally, it incorporates relevant scene observations, such as traffic signals, road topology and agent detections, through cross-attention mechanisms. This forms an efficient data-streaming architecture, that contrasts with the expensive, fixed-sequence input common in existing models. The architecture also offers the distinct advantage of generating occupancy predictions through localized querying based on a point-of-interest, as opposed to generating fixed-size occupancy images that render potentially irrelevant regions. Copyright © 2023, The Authors. All rights reserved.

关键词： Computer vision

来源：评论

学校读者我要写书评

暂无评论

Pushing AI to wireless network edge: an overview on integrated sensing, communication, and computation towards 6G

引用

Science China(Information Sciences) 2023年第3期66卷 7-25页

作者： Guangxu ZHU Zhonghao LYU Xiang JIAO Peixi LIU Mingzhe CHEN Jie XU Shuguang CUI Ping ZHANG Shenzhen Research Institute of Big Data Future Network of Intelligence Institute (FNii) The Chinese University of Hong Kong (Shenzhen) School of Science and Engineering (SSE) The Chinese University of Hong Kong (Shenzhen) State Key Laboratory of Advanced Optical Communication Systems and Networks School of Electronics Peking University Department of Electrical and Computer Engineering and Institute for Data Science and Computing University of Miami Peng Cheng Laboratory State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications

Pushing artificial intelligence(AI) from central cloud to network edge has reached board consensus in both industry and academia for materializing the vision of artificial intelligence of things(AIoT) in the sixth-generation(6G) era. This gives rise to an emerging research area known as edge intelligence, which concerns the distillation of human-like intelligence from the vast amount of data scattered at the wireless network edge. Typically, realizing edge intelligence corresponds to the processes of sensing, communication,and computation, which are coupled ingredients for data generation, exchanging, and processing, ***, conventional wireless networks design the three mentioned ingredients separately in a task-agnostic manner, which leads to difficulties in accommodating the stringent demands of ultra-low latency, ultra-high reliability, and high capacity in emerging AI applications like auto-driving and metaverse. This thus prompts a new design paradigm of seamlessly integrated sensing, communication, and computation(ISCC) in a taskoriented manner, which comprehensively accounts for the use of the data in downstream AI tasks. In view of its growing interest, this study provides a timely overview of ISCC for edge intelligence by introducing its basic concept, design challenges, and enabling techniques, surveying the state-of-the-art advancements, and shedding light on the road ahead.

关键词： sixth-generation (6G) edge intelligence artificial intelligence of things (AIoT) integrated sensing communication and computation (ISCC)

来源：评论

学校读者我要写书评

暂无评论

Vehicle Routing Problem Considering Electric Vehicle Battery Consumption

Vehicle Routing Problem Considering Electric Vehicle Battery...

引用

electrical, Computer, Communications and Mechatronics engineering (ICECCME), International Conference on

作者： Nikica Perić Slaven Begović Mohan Muniappan Vinko Lešić Faculty of Electrical Engineering and Computing Laboratory for Renewable Energy Systems University of Zagreb Zagreb Croatia

Delivery transport is one of the main drives of the global economy and a major oil consumer. Calculation of optimal routes is an everlasting effort to increase the efficiency of the sector. With the rising trend of electric vehicles, the sector is exploring this revolutionary alternative. This paper presents a vehicle routing problem optimized with the information of delivery-dependent battery consumption, depending on road conditions and cargo. A mathematical model of the vehicle is derived, encompassing the information of mechanical and electrical aspects. Further simplification is introduced with statistical analysis of typical driving cycles in the considered environments. Finally, the on-site road conditions are exploited to have a simple, yet sufficiently accurate information on the battery capacity at every point of the delivery process. The predictive vehicle routing algorithm is applied for the realistic case study and a comparison of electric and conventional diesel vehicles is performed, additionally compared to the current retail data provider. Results show promising monetary costs reductions of up to 80% in various cases, outlying the potential of the new technology and optimization algorithms.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Quantitative Analysis of Regression-Based Temperature Dynamics Models for Households with A/C Units Subject to Unknown Disturbances

Quantitative Analysis of Regression-Based Temperature Dynami...

引用

Human-Computer Interaction, Optimization and Robotic Applications (HORA), International Congress on

作者： Nikola Hure Mario Vašak University of Zagreb Faculty of Electrical Engineering and Computing Laboratory for Renewable Energy Systems Zagreb Croatia

This paper focuses on the identification of thermodynamic models for temperature prediction in households. The proposed temperature dynamics model falls under the class of Linear Time-Invariant (LTI) models, making it suitable for model predictive control synthesis. However, the presence of significant and variable thermal disturbances in households adds complexity to the identification process. The performance of various prediction error methods, such as ARX, ARARMAX, and BJ models, along with simplified models incorporating persistent disturbance excitation, is analyzed. The findings highlight the substantial impact of unknown disturbances on temperature predictions, emphasizing the crucial need for accurate prediction of these disturbances for effective household heating and cooling planning. The identification and evaluation of model performance measures are conducted using two months of experimental data collected from five households. This study contributes to understanding of the significance of addressing unknown disturbances and variability in thermodynamic model identification for temperature prediction.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Incorporating a High-speed Camera in the Lightning Current Measurement System for Wind Turbines

Incorporating a High-speed Camera in the Lightning Current M...

引用

International Colloquium on Smart Grid Metrology (SmaGriMet)

作者： Franjo Vukovic Viktor Milardic Bozidar Filipovic-Grcic Nina Stipetic Bojan Franc Dominik Milos Department of Energy and Power Systems Faculty of Electrical Engineering and Computing University of Zagreb Zagreb Croatia

A lightning current measurement system for wind turbines was developed and tested in the High Voltage Laboratory at the Faculty of electrical engineering and computing, University of Zagreb, to determine actual distributions of lightning current parameters in a particular wind turbine area. The output of the measurement system is a waveform of lightning current. The measurement system consists of two current sensors with different frequency and amplitude ranges and a real-time controller with two digitizers and a GPS module. The current sensors are Rogowski coils connected to integrators with coaxial cables. The prototype was installed on a wind turbine in a high winter lightning activity area. The current stage of development involves integrating the high-speed camera into the prototype measurement system to obtain the visual confirmation of the waveform measurements and help categorize lightning strikes into downward or upward strikes. The camera control application was developed in LabVIEW. The camera was set to trigger by the real-time controller. Also, a redundant trigger occurs when there is a change in motion in the camera frame in case the prototype measurement system fails to register a lightning strike and therefore fails to send a trigger signal to the camera. The camera control application was tested in the High Voltage Laboratory by recording a flashover in a spark gap and a flashover on the insulator.

关键词： Current measurement Lightning Prototypes High-voltage techniques Flashover Cameras Real-time systems

来源：评论

学校读者我要写书评

暂无评论

Boosting the Performance of Transformer Architectures for Semantic Textual Similarity

arXiv

引用

arXiv 2023年

作者： Rep, Ivan Čeperić, Vladimir Department of Electronics Microelectronics Computer and Intelligent Systems Faculty of Electrical Engineering and Computing Zagreb Croatia

Semantic textual similarity is the task of estimating the similarity between the meaning of two texts. In this paper, we fine-tune transformer architectures for semantic textual similarity on the Semantic Textual Similarity Benchmark by tuning the model partially and then end-to-end. We experiment with BERT, RoBERTa, and DeBERTaV3 cross-encoders by approaching the problem as a binary classification task or a regression task. We combine the outputs of the transformer models and use handmade features as inputs for boosting algorithms. Due to worse test set results coupled with improvements on the validation set, we experiment with different dataset splits to further investigate this occurrence. We also provide an error analysis, focused on the edges of the prediction range. © 2023, CC BY.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Structural loads reduction in wind turbines via set-based control

Structural loads reduction in wind turbines via set-based co...

引用

IEEE International Conference on Industrial Technology (ICIT)

作者： Ivan Grabić Mario Vašak Faculty of Electrical Engineering and Computing Laboratory for Renewable Energy Systems University of Zagreb Zagreb Croatia

The paper is focussed on set-based correction of the wind turbine controls in order to evade high mechanical loads on structural components of a wind turbine during transients. A set in the state space of the wind turbine model is constructed where all accounted mechanical loads are within the prescribed limits. The approach is universal in a way that mechanical loads on different components, which are expressed as functions of the wind turbine states and inputs, can be blended together in that set description. A procedure is proposed to determine a non-conservative polytopic under-approximation of the mentioned set. Following that, a controlled invariant set methodology is used to compute a controlled invariant subset of the polytopic set. It is applied in on-line controls for a minimalistic correction of outputs of any wind turbine control algorithm that assures all the observed mechanical loads on different components are within boundaries. The approach is illustrated for mechanical loads on the axle pin and controls in an operating area above the nominal wind speed.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Hybrid Machine Learning Model for Solar Energy Prediction 17

Hybrid Machine Learning Model for Solar Energy Prediction

引用

17th International Conference on Computer and Automation engineering, ICCAE 2025

作者： Phaengsomsri, Likhasit Xu, Ke Ma, Xuzhe Zhang, Yibo Mo, Huadong University of New South Wales School of Engineering & Technology Canberra Australia University of New South Wales School of Electrical Engineering and Telecommunications Sydney Australia University of Sydney Faculty of Engineering Sydney Australia University of New South Wales School of Systems and Computing Canberraa Australia

ISBN: (纸本)9798331533816

The increasing global energy demand necessitates sustainable energy solutions to mitigate the adverse effects of fossil fuels. Solar energy, harnessed through Photovoltaic (PV) systems, offers a viable alternative, but its variability poses integration challenges with the power grid. This study proposes a hybrid machine learning (ML) model combining Multi-Layer Perceptron Neural Network (MLPNN) for pattern recognition, Decision Tree (DT) for feature selection, and Bayesian Optimization for hyperparameter tuning to improve solar energy output prediction. Using historical meteorological data from a PV system in Canberra, Australia, the model identified solar irradiance as the most influential feature, with temperature, wind speed, and time contributing unique information. Results demonstrate that while the MLPNN achieved superior prediction accuracy, the hybrid model reduced training time by 30 % with a comparable RMSE, MSE, and R². These findings highlight a trade-off between accuracy and efficiency, suggesting the hybrid model for real-time applications and the MLPNN for accuracyfocused tasks. This research provides a practical framework for enhancing solar energy prediction systems. © 2025 IEEE.

关键词： Artificial Intelligence Machine Learning Solar

来源：评论

学校读者我要写书评

暂无评论

Underwater breadcrumb localization

Underwater breadcrumb localization

引用

OCEANS

作者： Vladimir Slošić Ðula Nađ Nikola Mišković Faculty of Electrical Engineering and Computing Laboratory for Underwater Systems and Technologies University of Zagreb Zagreb Croatia

In most cases, underwater localization implies a fusion of two or more sensors, usually inertial and acoustic ones. Acoustic localization demands the deployment of the acoustic baseline if precise measurement is needed. Using a fixed predeployed baseline limits the accuracy of location estimates outside certain operating areas defined by the baseline. In case an AUV or a diver has to be localized underwater, their trajectory is usually unknown, so it is a challenging task to deploy a baseline in such a manner that localization is accurate. In this paper, the sequential deployment of acoustic breadcrumb modems is suggested as the alternative to the conventional baseline deployment. This type of sequentially deployed acoustic baseline in combination with IMU is first simulated and later on tested in a constrained environment in shallow waters in order to show what are the main challenges, advantages and potential problems with this type of localization.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Aggregated Model of a Flexible Multi-Energy Distribution System for Coordinated TSO-DSO Fast Frequency Services Provision

Aggregated Model of a Flexible Multi-Energy Distribution Sys...

引用

International Conference on Smart Energy systems and Technologies (SEST)

作者： Bojana Barać Matej Krpan Tomislav Capuder Department of Energy and Power Systems Faculty of Electrical Engineering and Computing University of Zagreb Zagreb Croatia

The massive penetration of distributed energy resources (DER) encourages the provision of frequency regulation services at the distribution network level following large frequency disturbances occurring at the transmission system side. The availability of reduced-order or aggregated models will facilitate the coordination between transmission and distribution system operators. As there are no models that can capture the dynamics of diverse distributed technologies and their control systems by a unified single device model, this paper presents the single equivalent device based on the virtual synchronous machine (VSM) technology that can subsume the frequency response of different DER and loads. The validation of the proposed model is performed on the adjusted IEEE 9-bus system for different multi-energy systems. Simulated scenarios have demonstrated the accuracy of proposed aggregation with a maximum relative error of 0.1% as well as computational improvement of almost 100%.

关键词：

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：