检索结果-内蒙古大学图书馆

Transformer-CNN for small image object detection

Signal Processing: Image Communication 2024年 129卷

作者： Chen, Yan-Lin Lin, Chun-Liang Lin, Yu-Chen Chen, Tzu-Chun Department of Electrical Engineering National Chung Hsing University Taichung 40227 Taiwan Department of Automatic Control Engineering Feng Chia University Taichung 40724 Taiwan

Object recognition in computer vision technology has been a popular research field in recent years. Although the detection success rate of regular objects has achieved impressive results, small object detection (SOD) is still a challenging issue. In the Microsoft Common Objects in Context (MS COCO) public dataset, the detection rate of small objects is typically half that of regular-sized objects. The main reason is that small objects are often affected by multi-layer convolution and pooling, leading to insufficient details to distinguish them from the background or similar objects, resulting in poor recognition rates or even no results. This paper presents a network architecture, Transformer-CNN, that combines a self-attention mechanism-based transformer and a convolutional neural network (CNN) to improve the recognition rate of SOD. It captures global information through a transformer and uses the translation invariance and translation equivalence of CNN to maximize the retention of global and local features while improving the reliability and robustness of SOD. Our experiments show that the proposed model improves the small object recognition rate by 2∼5 % than the general transformer architectures. © 2024

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Construction and optimization of a computational holographic lithography system 8

Construction and optimization of a computational holographic...

引用

8th International Conference on Speckle Metrology, Speckle 2023

作者： Chen, Rui Lin, Angyi You, Jianfeng Lin, Huixiang Zhao, Wenqing Zhang, Fucai School of Control and Computer Engineering North China Electric Power University Baoding071003 China The Department of Electrical and Electronic Engineering Southern University of Science and Technology Shenzhen518055 China Engineering Research Center of Intelligent Computing for Complex Energy Systems Ministry of Education North China Electric Power University China

ISBN: (数字)9781510674592

ISBN: (纸本)9781510674585

Traditional lithography techniques are currently facing challenges, including high costs and the susceptibility of mask plates to damage. This research aims to elucidate the feasibility and technical constraints of a novel diffraction lithography approach. This method can reduce the size and cost of lithography machines, break through existing technological barriers, and offer new possibilities for the future development of lithography machines. We have built a basic computational holographic lithography system utilizing a 405nm laser and validated the system using holograms obtained with both classical GS algorithm and a newly proposed method. With the GS holograms, patterns with 700nm features have been successfully exposed on the photoresist. With our new algorithms, multiple holograms for the same target pattern can be generated with different settings in the algorithm. Sequential exposure of those holograms resulted in continuous lines 500 nm wide with reduced speckle artifacts, which was about a 1.4-time improvement over the result from the GS algorithm. Future research will focus on optimizing and enhancing the experimental system and the hologram calculation methods to reduce the line widths further and improve the exposure pattern quality. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Holograms

来源：评论

学校读者我要写书评

暂无评论

An Interacting Multiple Model Approach Based on Maximum Correntropy Student's T Filter

An Interacting Multiple Model Approach Based on Maximum Corr...

引用

IEEE/RSJ International Conference on Intelligent Robots and systems (IROS)

作者： Fethi Candan Aykut Beke Lyudmila Mihaylova Automatic Control and Systems Engineering University of Sheffield Sheffield United Kingdom Department of Control Systems Design ASELSAN Ankara Turkiye

This paper presents a novel approach called the Interacting Multiple Model (IMM)-based Maximum Correntropy Student's T Filter (MCStF), which addresses the challenges posed by non-Gaussian measurement noises. The MCStF demonstrates superior performance compared to the IMM algorithm based on Kalman Filters (KFs) in both simulation environments and real-time systems. The Crazyflie 2.0 nano Unmanned Air Vehicle (UAV) model is used in the simulation validation, and results from 3000 independent Monte Carlo runs are shown. After getting the simulation results under monotonously changed non-Gaussian distribution, their performance results have been compared to each other. The same scenario has been applied in the real-time system using Crazyflie 2.0. Next, results from real-time tests are presented in which the position of Crazyflie 2.0 is estimated online.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An Approach for Small Image Object Detection Based on Transformer-Cnn

SSRN

引用

SSRN 2024年

作者： Lin, Chun-Liang Chen, Yan-Lin Lin, Yu-Chen Department of Electrical Engineering National Chung Hsing University Taichung40227 Taiwan Department of Automatic Control Engineering Feng Chia University Taichung40724 Taiwan

Although the detection success rate of regular objects has achieved excellent results, small object detection (SOD) is still challenging in object recognition. The main reason is that small objects are often affected by multi-layer convolution and pooling leading to resulting in a lack of sufficient details to distinguish them from the background or similar objects leading to poor recognition rates or even no results. This paper presents a network architecture, named Transformer-CNN, that combines a self-attention mechanism-based transformer and convolutional neural network (CNN) to improve the recognition rate of SOD. It captures global information through a transformer and uses the translation invariance and translation equivalence of CNN to maximize the retention of global and local features while improving the reliability and robustness of SOD. Our experiments show that the proposed model improves the small object recognition rate by 2~5% than the general transformer architectures. © 2024, The Authors. All rights reserved.

关键词： Object recognition

来源：评论

学校读者我要写书评

暂无评论

Lyapunov-Inspired Deep Reinforcement Learning for Robot Navigation in Obstacle Environments

SSRN

引用

SSRN 2024年

作者： Kayacan, Erdal Ugurlu, Halil Ibrahim Redder, Adrian Department of Electrical and Computer Engineering Aarhus University Aarhus C Denmark Department of Automatic Control Paderborn University Paderborn Germany

After its success in games and simulated control tasks, deep reinforcement learning (DRL) has gained prominence in robotics for developing neural network-based planning and control algorithms. However, the inherent black-box nature of neural network controllers poses challenges in ensuring safety, prompting the need for innovative, explainable solutions. This research paper introduces a DRL algorithm design inspired by Lyapunov stability theory to solve safe robot navigation problems in the presence of obstacles. The robot planning problem is formulated as a state-space control problem with obstacle locations integrated into the state representation. To ensure a high probability of obstacle avoidance and, hence, safety, we introduce a reward-shaping strategy utilizing a Lyapunov function tailored for obstacle avoidance. Additionally, an early episode termination is employed to constrain exploration and add more valuable samples to the DRL training replay memory. Simulation results demonstrate the effectiveness of the method in achieving superior exploration and learning with higher mission completion rates while maintaining speeds closer to a desired target speed. Real-world experiments conducted with a quadrotor in a laboratory environment validate the applicability of our method. The findings underscore the potential of incorporating Lyapunov stability theory into DRL approaches for advancing safe and efficient robot navigation. © 2024, The Authors. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Exponentially convergent distributed Nash equilibrium seeking for constrained aggregative games

引用

Autonomous Intelligent systems 2022年第1期2卷 71-78页

作者： Shu Liang Peng Yi Yiguang Hong Kaixiang Peng Department of Control Science&Engineering Tongji UniversityShanghai200092China Shanghai Research Institute for Intelligent Autonomous Systems Shanghai201210China School of Automation and Electrical Engineering University of Science and Technology BeijingBeijing100083China

Distributed Nash equilibrium seeking of aggregative games is investigated and a continuous-time algorithm is *** algorithm is designed by virtue of projected gradient play dynamics and aggregation tracking dynamics,and is applicable to games with constrained strategy sets and weight-balanced communication *** key feature of our method is that the proposed projected dynamics achieves exponential convergence,whereas such convergence results are only obtained for non-projected dynamics in existing works on distributed optimization and equilibrium *** examples illustrate the effectiveness of our methods.

关键词： Distributed algorithms Aggregative games Projected gradient play Weight-balanced graph Exponential convergence

来源：评论

学校读者我要写书评

暂无评论

Adaptability analysis on the measurement device of UACV special machinery unlocking force

Adaptability analysis on the measurement device of UACV spec...

引用

2023 International Conference on Mechatronic engineering and Artificial Intelligence, MEAI 2023

作者： Yang, Mo Cai, Zizhuo Lin, Jing Li, Jiahao Ren, Haiyue Liu, Qiang School of Mechanical Engineering Nanjing University of Science and Technology 200 Xiaolingwei Street Nanjing210094 China Department of Automatic Control Systems Bauman Moscow State Technical University ul. Baumanskaya 2-ya 5/1 Moscow105005 Russia

ISBN: (纸本)9781510674608

During the unlocking process of the rotating-bolt locking special machinery head, there is a significant resistance. The aim is to select a motor for unmanned aerial combat vehicle (UACV) mounted external power special machinery drives, taking the special machinery unlocking force as the research subject. Adaptive analysis is conducted on the patented structure of the special machinery unlocking force measurement device previously designed. A virtual prototype model is established using the dynamic simulation software ADAMS to apply driving, constraints, and simulation parameters based on the model's principles. Addressing the phenomenon of the detachment of the guide rod slider from the tested bolt carrier during the operation of the special machinery unlocking force measurement device, it is hypothesized that the detachment time is related to the stiffness of the spring and the velocity of the lead screw slider. Through multiple comparative simulations, the correctness of the hypothesis is proven. It is demonstrated that increasing the velocity of the lead screw slider can completely eliminate the detachment time, whereas increasing the stiffness of the spring, although reducing the detachment time, still results in secondary detachment. Consequently, the conclusion is drawn that increasing the velocity of the lead screw slider is the way to resolve the detachment issue of the guide rod slider from the tested bolt carrier in the experiment. © 2024 SPIE.

关键词： Unmanned aerial vehicles (UAV)

来源：评论

学校读者我要写书评

暂无评论

Learn to Bid as a Price-Maker Wind Power Producer

arXiv

引用

arXiv 2025年

作者： Singhal, Shobhit Fochesato, Marta Aolaritei, Liviu Dörfler, Florian Automatic Control Laboratory ETH Zürich Switzerland Department of Electrical Engineering and Computer Sciences UC Berkeley United States

Wind power producers (WPPs) participating in short-term power markets face significant imbalance costs due to their non-dispatchable and variable production. While some WPPs have a large enough market share to influence prices with their bidding decisions, existing optimal bidding methods rarely account for this aspect. Price-maker approaches typically model bidding as a bilevel optimization problem, but these methods require complex market models, estimating other participants’ actions, and are computationally demanding. To address these challenges, we propose an online learning algorithm that leverages contextual information to optimize WPP bids in the price-maker setting. We formulate the strategic bidding problem as a contextual multi-armed bandit, ensuring provable regret minimization. The algorithm’s performance is evaluated against various benchmark strategies using a numerical simulation of the German day-ahead and real-time markets. Copyright © 2025, The Authors. All rights reserved.

关键词： Wind power

来源：评论

学校读者我要写书评

暂无评论

Intelligent Fault Diagnosis for Low-Voltage Power Network

Intelligent Fault Diagnosis for Low-Voltage Power Network

引用

2023 IEEE International Conference on Energy Technologies for Future Grids, ETFG 2023

作者： Anwar, Hamza M. Shafiullah, Md Abido, M.A. Kfupm Control & Instrumentation Engineering Department Dhahran31261 Saudi Arabia Electrical Engineering Department Dhahran31261 Saudi Arabia Kfupm Irc for Renewable Energy and Power Systems Dhahran31261 Saudi Arabia Dhahran31261 Saudi Arabia

ISBN: (纸本)9781665471640

Faults are common and potentially dangerous problems in low-voltage power system networks, e.g., distribution networks. Upon discovery, they must be dealt with immediately to prevent further damage to the distribution system and quickly return electricity to consumers. This paper proposes a hybrid fault diagnosis method combining a signal processing technique with deep learning models. Faults are applied in a four-node test distribution feeder developed using MATLAB and Simulink. Different measurement noise levels are implemented, along with varying load and fault parameters. The short-time Fourier transform (STFT) is used on the feeder's current signals as a feature extraction tool. These features are then used to train deep learning models to detect, classify, and locate faults. Various parameters of the models are varied to find the optimum ones, which are used to obtain the results. The findings show that the model performed exceptionally well in fault detection and classification and satisfactorily in fault location. © 2023 IEEE.

关键词： deep neural networks fault classification fault detection fault location MATLAB/SIMULINK Python Short-time Fourier Transform

来源：评论

学校读者我要写书评

暂无评论

Analysis of hybridization for the use of hydrogen for aircraft propulsion in an existing platform

Analysis of hybridization for the use of hydrogen for aircra...

引用

AIAA Aviation and Aeronautics Forum and Exposition, AIAA AVIATION Forum 2023

作者： de Frutos, Víctor M. Parra, Juan R. Esteban, Sergio Bordons, Carlos Systems Engineering and Automatic Control Escuela Técnica Superior de Ingeniería de Sevilla Universidad de Sevilla Sevilla Spain Aeroespace Engineering Department Escuela Técnica Superior de Ingeniería de Sevilla Universidad de Sevilla Sevilla Spain

ISBN: (纸本)9781624107047

This work presents a methodology to analyze the hybridization use of fuel cells, batteries, and combustion engines to supply the power needed for flight in all the phases. The analysis is performed on a case study analyzing the diagram Payload vs. Range for different degrees of hybridization and solutions as well as different technology levels of the main components of the propulsion system based on fuel cells. The penalty in the utilization of hydrogen with fuel cells in the aircraft is high. Several hybridization strategies are analyzed to comply with more demanding sizing requirements as take-off. The calculation methodology is applied for regional aircraft to evaluate the impact of the introduction of fuel cell as a propulsion system in an existing platform and extended to other aviation segments. The objective of the presented work is to determine if green propulsion with fuel cells is possible, and in which flight segments is more suitable, so it is analyzed the improvement needed in the fuel cell technologies and energy storage and also different A/C are analyzed. Finally, it is faced the impact in the design point of the aircraft for the introduction of this new technology. © 2023, American Institute of Aeronautics and Astronautics Inc, AIAA. All rights reserved.

关键词： Fuel cells

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：