检索结果-内蒙古大学图书馆

Causal-Transformer: Spatial-temporal causal attention-based transformer for time series prediction

IFAC-PapersOnLine 2024年第14期58卷 79-84页

作者： Yaqi Zhu Fan Yang Andrei Torgashov Department of Automation Beijing National Research Center for Information Science and Technology Tsinghua University Beijing 100084 China Process Control Laboratory Institute of Automation and Control Processes Far-Eastern Branch Russian Academy of Sciences Vladivostok 690041 Russia

Real-time monitoring and accurate prediction of key variables are indispensable to ensure industrial production activities proceed as expected. With the increase in measurement data volume and the improvement of hardware computing power, the Transformer and its variants, due to their excellent capability in extracting global dependencies, are playing an increasingly important role among deep learning-based multidimensional time series prediction models. In addition, from the perspective of causality, cause variables contain parts of information in effect variables and can reduce the uncertainty of effect variables, which is beneficial for prediction. However, there has been relatively limited research on combining the Transformer and causal feature analysis. To fully use both advantages, this paper introduces the Causal-Transformer (CT) model, which utilizes semi-orthogonal projection to extract causal features from multiple input variables. A multi-head spatial-temporal causal attention mechanism is designed in the encoder block based on the classical Transformer model to simultaneously reduce feature dimensions and extract implicit causal features in both the temporal and spatial dimensions. The CT also utilizes the Granger causality analysis to select the causal teaching indicators of target variables to provide stable assistance by injecting explicit causality into the inputs of the decoder block. By leveraging more condensed and independent causal features, the CT possesses inherent advantages in predicting time series variables. Case study results show that the CT model outperforms the other models on the diesel refinery dataset, especially with a reduction of 46.0% and 30.4% in MSE towards the classic Transformer and informer in five-step prediction.

关键词： Machine learning Time series modelling Transformer Causal analysis Attention machenism

来源：评论

学校读者我要写书评

暂无评论

The Use of Virtual Interactive Stands in the Educational process in Higher Education 4

The Use of Virtual Interactive Stands in the Educational Pro...

引用

4th International Conference on Technology Enhanced Learning in Higher Education, TELE 2024

作者： Ubaydullayeva, Shakhnoza Gulyamova, Zarina Tadjiyeva, Gavhar Kadirova, Nilufar Kusanova, Ugiloy National Research University Department 'Automation and Control of Technology Process Tashkent Institute of Irrigation and Agricultural Mechanization Engineers Tashkent Uzbekistan Tashkent State Pedagogical University Named after Nizami Department of Theory and Methodology of Teaching English Language Tashkent Uzbekistan

ISBN: (纸本)9798350353532

Information and communication technologies have become an integral part of modern society in Uzbekistan, as they are used in all spheres of human life: economic, political and social. Uzbekistan is one of the countries where digital technologies are rapidly developing, because of which significant results have been achieved in education. The expansion of distance education will improve the quality of education, because more students who live in rural areas are covered. This will allow you to receive high-quality content (video, audio, and platform);use the most up-to-date technologies and solutions: online services for the educational process, gamification, a virtual library, online proctoring and practical online classes. Interactive learning is a learning process using information and communication technologies. Although distance learning is not a substitute for university attendance, it is an excellent learning tool for consolidating knowledge and obtaining additional facts and information. Sometimes homework is not enough for proper understanding and assimilation of educational material. Proper understanding of information is useful for practical subjects. The Internet provides a wide range of such additional features. Interactive stands enhance both general and professional knowledge;have a motivating effect on learning in a playful way. The National Research University in Uzbekistan, together with the teachers of the Tashkent State Pedagogical University, is actively working on the introduction of virtual interactive educational stands into the educational process, including distance education. The computer program provides for the implementation of training lessons in the section 'Logical foundations of computing systems', namely 'Representation of functions of the algebra of logic', 'Minimization of logical functions by the Veitch diagram method', 'Minimization of logical functions by the Quine method' © 2024 IEEE.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

A Norm Optimal Iterative Learning control Approach with Efficient Computation 13

A Norm Optimal Iterative Learning Control Approach with Effi...

引用

13th IEEE Data Driven control and Learning Systems Conference, DDCLS 2024

作者： Gao, Luyuan Zhuang, Zhihe Tao, Hongfeng Chen, Yiyang Paszke, Wojciech Jiangnan University Key Laboratory of Advanced Process Control for Light Industry of Ministry of Education Wuxi214122 China School of Mechanical and Electrical Engineering Soochow University Suzhou215137 China Institute of Automation University of Zielona Gora Zielona Gora Poland

ISBN: (纸本)9798350361674

Iterative learning control (ILC) is excellently applied to output tracking for systems that perform repetitive tasks. Classical norm optimal ILC (NOILC) approach yields optimal control inputs by minimizing a quadratic cost function. However, this approach requires the introduction of supervectors, i.e., lifting technique, for system description, which raises a high computational load problem. The contribution of this paper is to propose a non-lifted NOILC (N-NOILC) approach by redesigning a cost function that does not involve the lifting technique, achieving the goal of efficient computation. Therefore, the proposed approach is particularly suitable for large tracking tasks with long trial duration and high sampling frequency. Moreover, considering the model uncertainty in practice, robust convergence is analyzed. Numerical simulations are shown to demonstrate the properties and advantages of the proposed approach. © 2024 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Fast Integral Terminal Sliding Mode control with Disturbance Compensation for Permanent Magnet Synchronous Motors

Fast Integral Terminal Sliding Mode Control with Disturbance...

引用

第43届中国控制会议

作者： Xiaomin Mao Chenglin Liu Liang Shan Key Laboratory of Advanced Process Control for Light Industry of Ministry of Education Institute of AutomationJiangnan University School of Automation Nanjing University of Science and Technology

ISBN: (数字)9789887581581

ISBN: (纸本)9798350366907

To enhance the robustness of the permanent magnet synchronous motor(PMSM) speed control system, and reduce the response time of the control system, a composite terminal sliding mode control strategy is proposed. Firstly, a integral terminal sliding surface and a fast reaching law are designed based on the mathematical model of the PMSM to avoid singularity issues,enhance quickness, and reduce chattering. Then, the stability and finite-time convergence of the control system are proved by constructing a Lyapunov function. Considering the system's susceptibility to disturbances, the lumped disturbance is considered in the mathematical model. An extended state observer(ESO) is designed to observe the lumped disturbance and compensate for the lumped disturbance to the controller. Finally, through simulation and experimentation, it is demonstrated that the designed controller exhibits advantages such as fast response, no overshoot, and strong robustness.

关键词： Permanent Magnet Synchronous Motor Speed control System Terminal Sliding-mode control Extended State Observer

来源：评论

学校读者我要写书评

暂无评论

H∞ Tracking control of Tow Time Scale Linear System Based on Off-Policy Reinforcement Learning

H∞ Tracking Control of Tow Time Scale Linear System Based o...

引用

Data Driven control and Learning Systems (DDCLS)

作者： MinYin Shen Fei Liu Key Laboratory of Advanced Process Control for Light Industry (Ministry of Education) Institute of Automation Jiangnan University Wuxi China

ISBN: (数字)9798350361674

ISBN: (纸本)9798350361681

In this paper, by integrating singular perturbation method and off-policy reinforcement learning, a new approach to solving the H∞ tracking control problem is presented for linear time-invariant two time scales systems. A composite control scheme is designed through decomposing the two time scales system into fast and slow subsystems. For the slow subsystem, a model-free H∞ tracking controller is designed based on off-policy reinforcement learning to accomplish system tracking of the desired trajectory while effectively suppressing disturbances, without knowledge of the dynamics. Meanwhile, for the fast subsystem, a model-based offline controller is adopted. To better tracking performance, an exponential discount factor is introduced to weight the cost of input and tracking errors. By analysis and simulation, the effectiveness and superior tracking performance of proposed method has been verified.

关键词： Linear systems Learning systems Costs Perturbation methods Design methodology Reinforcement learning control systems

来源：评论

学校读者我要写书评

暂无评论

Fixed-time Funnel Sliding Mode control for Dual-motor Driving System

Fixed-time Funnel Sliding Mode Control for Dual-motor Drivin...

引用

Fully Actuated System Theory and Applications (CFASTA), Conference on

作者： Jia-Meng Wu Cheng-Lin Liu Key Laboratory of Advanced Process Control for Light Industry (Ministry of Education) Institute of Automation Jiangnan University Wuxi China

ISBN: (数字)9798350373691

ISBN: (纸本)9798350373707

A fixed-time funnel sliding mode control strategy is developed for a dual-motor driving system, which ensures the system converges within a fixed time. Firstly, the coupling error is introduced by combining the reference tracking process with the motor synchronization process. By using a selected funnel function, the system can achieve satisfactory transient performance. Secondly, the concept of multi-surface sliding modes is designed to deal with the tracking issue so that the system can converge in a fixed-time regardless of the motor’s initial state. Finally, the effectiveness of this strategy is demonstrated through a simulation example.

关键词： Couplings Motors Synchronization Transient analysis Sliding mode control Convergence

来源：评论

学校读者我要写书评

暂无评论

Model-Free $H_{2}/H_{\infty}$ Predictive control for Discrete-Time System via Q-Learning

Model-Free $H_{2}/H_{\infty}$ Predictive Control for Discret...

引用

Data Driven control and Learning Systems (DDCLS)

作者： Yihong Lin Peng He Haiying Wan Zhuangyu Liu Xiaoli Luan Fei Liu Key Laboratory of Advanced Process Control for Light Industry Ministry of Education Institute of Automation Jiangnan University Wuxi China

ISBN: (数字)9798350361674

ISBN: (纸本)9798350361681

This paper presents a model-free $H_{2}/H_{\infty}$ Q-learning predictive control strategy for linear discrete-time systems. To design predictive controller with the system measured states, a policy iteration solution algorithm is employed to approximate the control inputs. Specifically, the developed algorithm is formulated in the form of linear matrix inequalities, designed to stabilize the system with mixed $H_{2}/H_{\infty}$ control performance. Additionally, to improve robust performance under system disturbance variations, we introduce the receding horizon optimization into the Q-learning based predictive control. Finally, simulation results demonstrate the effectiveness of the proposed approach.

关键词： Discrete-time systems Q-learning Simulation Heuristic algorithms Predictive models Prediction algorithms Approximation algorithms

来源：评论

学校读者我要写书评

暂无评论

Integral Adaptive Sliding Mode control for Vehicle Platoon with Disturbance Observer

Integral Adaptive Sliding Mode Control for Vehicle Platoon w...

引用

Fully Actuated System Theory and Applications (CFASTA), Conference on

作者： Xin Yang Cheng-Lin Liu Key Laboratory of Advanced Process Control for Light Industry (Ministry of Education) Institute of Automation Jiangnan University Wuxi China

ISBN: (数字)9798350373691

ISBN: (纸本)9798350373707

This article proposes an integral adaptive sliding mode control (IASMC) strategy that combines adaptive reaching laws and disturbance observers (DO) for vehicle platoon system subject to unknown disturbance. The objective is to ensure the string stability and make the spacing error between any front and rear vehicles in the platoon be zero, which means that the vehicles are driven at the expected spacing. To achieve it, the bidirectional (BD) topology and the constant time headway (CTH) policy are employed. Then, an adaptive smoothed sliding mode reaching law is constructed to eliminate the chattering issue. The unknown disturbances can be estimated by the designed DO. Meanwhile, the Lyapunov method is applied to the analysis and proof of the stability of the designed IASMC strategy, and the Laplace transform is used in the proof of the string stability. Finally, based on simulation, it can be seen that the availability and superiority of the designed strategy can be verified.

关键词： Adaptive systems Laplace equations Stability analysis Disturbance observers Topology Sliding mode control Lyapunov methods

来源：评论

学校读者我要写书评

暂无评论

Use of Elements and Algorithms of Intelligent Support in the automation of Technologies for control and Quality Management of Bulk Materials

Use of Elements and Algorithms of Intelligent Support in the...

引用

2022 International Conference Automatics and Informatics, ICAI 2022

作者： Kalandarov, Palvan I. Ubaydulayeva, Shakhnoza R. Gaziyeva, Rano T. Nikolov, Nikola N. Alexsandrova, Mariela I. National Research University Tashkent Institute of Irrigation and Agricultural Mechanization Engineers Department Automation and Control of Technology Process in Production Tashkent Uzbekistan Technical University of Varna Faculty of Computer Science and Automation Department of Automation Varna9010 Bulgaria

ISBN: (数字)9781665476256

ISBN: (纸本)9781665476256

The article deals with the choice of criteria for efficiency, applicability in measuring technology, analytical and construction, on the basis of which a methodology for assessing the quality of structures of measuring devices of moisture of materials of the agro-industrial complex is proposed. The standard deviation of the random error characterizing the accuracy of measurements is shown. A complex efficiency criterion is discussed, including the dominant particular parameters: accuracy, reliability, cost, allowing to solve the problem of optimal design of humidity control devices for the materials under consideration. As an optimal option, a differential measuring device is analyzed, which practically eliminates random and systematic errors and makes it possible to design a multiparameter measurement method that considers the features of a heterogeneous environment, such as bulk grain materials of the agro-industrial complex © 2022 IEEE.

关键词： Random errors

来源：评论

学校读者我要写书评

暂无评论

Predictive Battery Cooling in Heavy-Duty Fuel Cell Electric Vehicles 10th

Predictive Battery Cooling in Heavy-Duty Fuel Cell Electric ...

引用

10th IFAC Symposium on Advances in Automotive control, AAC 2022

作者： Büyüker, Banu Ç. Ferrara, Alessandro Hametner, Christoph Institute of Mechanics and Mechatronics Division of Process Control and Automation TU Wien Vienna Austria Christian Doppler Laboratory for Innovative Control and Monitoring of Automotive Powertrain Systems TU Wien Vienna Austria

In electric vehicles, it is essential to prevent battery overheating due to excessive ohmic losses or inadequate cooling. Indeed, the temperature of battery systems significantly impacts their performance, lifetime, and safety. This paper proposes a predictive cooling optimization method for the battery thermal management system of heavy-duty fuel cell electric vehicles. The predictive cooling strategy is based on a model predictive control (MPC) formulation to maintain the battery temperature in its optimal range (to increase efficiency) and avoid high-temperature peaks (to increase lifetime and safety). The predictive thermal management relies on the ohmic losses forecast provided by a predictive energy management system. Simulations of a real-world driving cycle validate the proposed MPC and assess the impact of the predictive horizon length, which is critical for thermal management performance. The comparison against a simple hysteresis control strategy highlights the significant benefits of the proposed MPC for higher battery efficiency and lifetime. Copyright © 2022 The Authors.

关键词： Model predictive control

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：