检索结果-内蒙古大学图书馆

63rd IEEE Conference on Decision and control, CDC 2024

作者： Wang, Jiwei Baldi, Simone Yu, Wenwu Yin, Xiang Southeast University School of Cyber Science and Engineering Nanjing210096 China Southeast University School of Mathematics Nanjing210096 China Shanghai Jiao Tong University Department of Automation Key Laboratory of System Control and Information Processing Shanghai200240 China

ISBN: (纸本)9798350316339

Fault diagnosis and prognosis in discrete event systems are studied in the scenario where the observations are possibly received with delay. To address this scenario, two conditions for diagnosis and prognosis with delayed observations are proposed, where we show that the state-of-the-art notion of prognosability must be revised to avoid conservativeness. Diagnosability and prognosability conditions are then verified by introducing a delay observer and a new verification function. Theoretical analysis indicates the effectiveness of the verification method for fault diagnosis and prognosis in the system. © 2024 IEEE.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multi-agent Deep Reinforcement Learning Based Multi-objective Charging control for Electric Vehicle Charging Station 7th

Multi-agent Deep Reinforcement Learning Based Multi-objectiv...

引用

7th Chinese Conference on Swarm Intelligence and Cooperative control, CCSICC 2023

作者： Gong, Jiabao Fu, Weiming Kang, Yu Qin, Jiahu Xiao, Feng Department of Automation University of Science and Technology of China Hefei230027 China State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources School of Control and Computer Engineering North China Electric Power University Beijing102206 China

ISBN: (纸本)9789819733279

This paper proposes a multi-agent deep reinforcement learning (MADRL) based algorithm for charging control of multiple electric vehicles (EVs) in an electric vehicle charging station (EVCS) with dynamic operations. By taking the uncertainties of EV users’ behaviors and charging prices into consideration, a Markov game with multiple objectives is first established, where the individual objective of reducing EV charging costs and the collective objective of mitigating charging load fluctuations are both studied. Then, a multi-agent actor-double-critic (MAADC) algorithm with a unified advantage actor-critic framework is proposed, where two advantage functions are designed as the local and global critics, and the integrated actor is employed to balance the learning towards the individual and collective objectives. The simulation results obtained in a delicately built EVCS environment show the effectiveness of MAADC. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Digital-Twin-Based Clock Synchronization in Industrial IoT systems 13

Digital-Twin-Based Clock Synchronization in Industrial IoT S...

引用

13th IEEE International Conference on CYBER Technology in automation, control, and Intelligent systems, CYBER 2023

作者： Cheng, Ting Liu, Shuai Xu, Qimin Zhu, Shanying Department of Automation Shanghai Jiao Tong University Shanghai200240 China Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai200240 China Shanghai Engineering Research Center of Intelligent Control and Management Shanghai200240 China School of Control Science and Engineering Shandong University Jinan250061 China

ISBN: (纸本)9798350315196

Close and efficient cooperation between devices in the Industrial Internet of Things (IIoT) requires precise clock synchronization as a prerequisite. The uncertainty of IIoT networks and the complexity of industrial environments have led to less precise clock synchronization achieved by traditional algorithms and higher network resource consumption. This paper proposes a digital-twin-based clock synchronization scheme for IIoT systems. By building the digital model for each local clock to predict their clock skew, the information interaction required to achieve synchronization is reduced. In order to model the clock accurately, this paper proposes to use the extended state Kalman filter to handle uncertain dynamics and unknown measurement bias to improve the estimation accuracy of clock skew further. The simulation results show that the proposed algorithm can achieve high- precision clock synchronization while reducing the consumption of network resources. © 2023 IEEE.

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Influence of the Kalman Filtering Initial Setting on SINS Initial Alignment and an Improved Method for the Alignment of Large Misalignment Angles 30

Influence of the Kalman Filtering Initial Setting on SINS In...

引用

30th Saint Petersburg International Conference on Integrated Navigation systems, ICINS 2023

作者： Ning, Xiaoge Huang, Jixun Li, Jianxun Shanghai Jiao Tong University School of Electronic Information and Electrical Engineering Department of Automation Shanghai China Beijing Aerospace Times Optical-Electronic Co. Ltd Beijing China Beijing Institute of Aerospace Control Devices Beijing China

ISBN: (纸本)9785919950950

This paper analyzes the instability of the filtering system of the initial alignment under a quasi-static base and points out the great influence of Kalman filtering initial value setting on the filtering results. This is instructive for many conditions under initial alignment and integrated navigation. Meanwhile, an improved method for the alignment of large misalignment angles is proposed by replacing the body-frame bias state with the geographic-frame bias state. Simulation experiments show that it can improve the horizontal alignment accuracy by 0.1°and vertical alignment accuracy by 1.6°. © 2023 Concern CSRI Elektropribor, JSC.

关键词： Alignment

来源：评论

学校读者我要写书评

暂无评论

Adversarially Robust Reinforcement Learning for Urban Traffic Signal control 5

Adversarially Robust Reinforcement Learning for Urban Traffi...

引用

5th IEEE International Conference on Technology, Informatics, Management, engineering and Environment, TIME-E 2024

作者： Hamzah, Muhammad Shafly Joelianto, Endra Sutarto, Herman Yoseph Putri, Adiyana Institut Teknologi Bandung Instrumentation and Control Master Program Bandung Indonesia Institut Teknologi Bandung Instrumentation Control and Automation Research Group Bandung Indonesia PT. Pusat Riset Energi Department of Intelligent System Bandung Indonesia Institut Teknologi Bandung Doctoral Program of Engineering Physics Bandung Indonesia

ISBN: (纸本)9798350364743

Urban traffic congestion presents critical challenges, resulting in delays, environmental pollution, and significant economic costs. While traditional traffic control methods, such as fixed-time and adaptive systems, have been widely implemented, they often fail in unpredictable conditions, particularly when confronted with noisy or inaccurate data. This research proposes a robust reinforcement learning (RL) solution for traffic signal control using the RADIAL-DQN (Robust Adversarial Loss DQN) algorithm. Integrating adversarial training enhances the system's robustness to real-world uncertainties. Evaluations conducted using a SUMO traffic simulation show that the robust agent produced an increase of only 60.71% in queue length under a projected gradient descent (PGD) adversarial attack, outperforming standard RL agents, which experienced an 89% increase. These results demonstrate the potential of robust RL to improve both the reliability and scalability of urban traffic control systems. © 2024 IEEE.

关键词： Traffic signals

来源：评论

学校读者我要写书评

暂无评论

INTRODUCING FEATURE MATCHING AND SMOOTHING LOSS: A NOVEL GAN FOR IMPROVED SAMPLE GENERATION 13

INTRODUCING FEATURE MATCHING AND SMOOTHING LOSS: A NOVEL GAN...

引用

13th International Conference on Quality, Reliability, Risk, Maintenance, and Safety engineering, QR2MSE 2023

作者： Yang, Junfeng Liu, Fengchang Mi, Jinhua Yin, Shengjie Bai, Libing School of Automation Engineering University of Electronic Science and Technology of China Chengdu China Center for System Reliability and Safety University of Electronic Science and Technology of China Chengdu China Department of SRM Design Shanghai Space Propulsion Technology Research Institute Shanghai China

ISBN: (纸本)9781839539848

Fault diagnosis of planetary gearboxes constitutes a vital element of rotating machinery. In recent years, fault diagnosis techniques for planetary gearboxes predicated on deep learning have undergone extensive research. However, in scenarios with limited samples, these deep learning methodologies can grapple with data scarcity issues, negatively impacting the model's generalization capacity and diagnostic performance. To mitigate this, Generative Adversarial Networks (GANs) have been integrated into planetary gearbox fault diagnosis to fabricate synthetic data, thereby augmenting the training dataset. Nevertheless, GANs face challenges in effectively learning local feature information from samples, leading to substantial disparities in local features between the generated data and the original data. Furthermore, the sampling of fault vibration signals from planetary gearboxes often exhibits a high frequency, with relatively smooth signal transitions, while the vibration signal samples generated by GANs often fluctuate excessively, demonstrating significant differences from the original vibration signals. To address this issue, this study introduces an enhanced GAN network that incorporates a feature attention matching mechanism in the computation of the loss function. This mechanism aims to amplify the feature similarity between the generated samples and local samples. Concurrently, a smoothness loss is introduced into the loss function to ensure the changes in the generated signal samples are more gradual. Experimental results indicate that this proposed method can generate superior quality samples, and the generated samples can achieve high accuracy on the WD-CNN diagnostic network. © The Institution of engineering & Technology 2023.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Weak-Model Based Reinforcement Learning control Strategy of Aircraft Attitude system

Weak-Model Based Reinforcement Learning Control Strategy of ...

引用

International Conference on Guidance, Navigation and control, ICGNC 2020

作者： Huang, Dingcui Hu, Jiangping Peng, Zhinan Hao, Mingrui School of Automation Engineering University of Electronic Science and Technology of China Chengdu611731 China Science and Technology on Complex System Control and Intelligent Agent Cooperation Laboratory Beijing100074 China

ISBN: (纸本)9789811581540

Intelligent aircraft control is an important strategic development trend. Traditional aircraft control algorithms have a strong dependence on the system models, and it is difficult to cope with the increasingly complex battlefield environment for intelligent aircrafts. In this paper, an online intelligent controller design based on a weak-model reinforcement learning technology is proposed to solve stabilization problem of an aircraft. The control problem is firstly converted to an optimal control problem, then an adaptive dynamic programming (ADP) technology is applied for computing the corresponding nonlinear Hamilton-Jacobi-Bellman equation (HJB) equation. Then, actor-critic neural networks structures are established to learn the optimal controller online where only partial information of the aircraft dynamics is required. The proposed control strategy enables the aircraft to adjust and update its control strategy on-line according to the actual mission targets and environments, such that the requirements of autonomous learning and intelligent operation can be realized. Finally, simulation example is offered to validate the effectiveness of the proposed weak-model based controller design. © 2022, The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Adaptive Deep Reinforcement Learning for Critical Boundary Scenario Generation 43

Adaptive Deep Reinforcement Learning for Critical Boundary S...

引用

43rd Chinese control Conference, CCC 2024

作者： Zhou, Junjie Wang, Lin Wang, Xiaofan Meng, Qiang Shanghai Jiao Tong University School of Electronic Information and Electrical Engineering Shanghai202400 China Ministry of Education of China Key Laboratory of System Control and Information Process Shanghai202400 China National University of Singapore Department of Civil and Environmental Engineering Singapore117576 Singapore Shanghai Institute of Technology School of Electrical and Electronic Engineering Shanghai202400 China

ISBN: (纸本)9789887581581

Critical boundary scenarios play a vital role in the comprehensive testing of autonomous vehicles, as they can significantly expedite testing processes and reduce economic and time costs. To tackle the intricate, multidimensional, and diverse nature of intelligent driving scenarios, this paper proposes an innovative bi-level adaptive deep reinforcement learning (BADRL) approach aimed at generating authentic and diverse critical boundary scenarios. Leveraging naturalistic driving data, background agents are trained to exhibit naturalistic and realistic driving behaviors using a neural-based naturalistic driving behavior model. To address the intricacies of multi-interaction, high-dimensional environments, the scenario complexity model is introduced to evaluate the relative complexity between the traffic environment and the tested autonomous vehicle in real time. By integrating the scenario complexity model, the naturalistic driving behavior learning, intelligent driving testing, and critical boundary scenario generation are concatenated together to form a closed loop. BADRL facilitates the upper-level neural network in learning to adaptively increase the complexity of test scenarios, which are then fed into lower-level models to optimize the behavior of principal traffic participants, thereby generating naturalistic and critical boundary test scenarios. Extensive simulations were carried out to verify the efficacy of the BADRL technique in complex intersection environments. Results indicate that the BADRL approach boosts the efficiency of critical boundary scenario generation by approximately 10% compared to the state-of-the-art method. © 2024 Technical Committee on control Theory, Chinese Association of automation.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Data-driven Fault Diagnosis Method Based on Fusion of Multi-direction and Multi-source Signals 5

Data-driven Fault Diagnosis Method Based on Fusion of Multi-...

引用

5th International Conference on Intelligent Autonomous systems, ICoIAS 2022

作者： Yang, Yang Wang, Junda Li, Ning Shanghai Jiaotong University Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai Engineering Research Center of Intelligent Control and Management Department of Automation Shanghai China

ISBN: (数字)9781665498388

ISBN: (纸本)9781665498388

As a pivotal mechanical component of turbine drive-chain, gearbox suffers from various accidents and fault diagnosis could remarkably enhance its reliability and maintain running safety and lifetime. Vibration-based analysis on single-channel is mainly adopted in previous diagnostic, which cannot catch full vibration inner information in gearbox sensors and bring model performance down. In this paper, a novel method based on the fusion of multi-direction and multi-source signals and the convolutional neural network (MDMS-CNN) is presented. This method can obtain multi-dimensional feature maps from MDMS and preserve the time correlation among different signal sequences. First, features from MDMS are extracted via a corrosion conversion method that converts signals into two-dimensional (2-D) gray images. Then, homologous images are spliced into same channels, while heterogeneous images are superimposed to different channels. Finally, fusion features are input into CNN for fault classification. In experiments, the method is validated on a gearbox diagnosis case. Results have shown that the proposed method achieves 99.86% recognition accuracy and faster convergence speed compared to others. © 2022 IEEE.

关键词： Failure analysis

来源：评论

学校读者我要写书评

暂无评论

Effective distributed algorithm for solving linear matrix equations

引用

Science China(Information Sciences) 2023年第8期66卷 265-266页

作者： Songsong CHENG Jinlong LEI Xianlin ZENG Yuan FAN Yiguang HONG Anhui Engineering Laboratory of Human-Robot Integration System and Intelligent Equipment School of Electrical Engineering and Automation Anhui University Department of Control Science and Engineering Tongji University Key Laboratory of Intelligent Control and Decision of Complex Systems School of AutomationBeijing Institute of Technology

Dear editor,Solving linear matrix equations is a basic and important problem in many fields such as the computation of generalized inverses of matrices and(generalized) Sylvester equations. Also, the linear algebraic ...

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：