检索结果-内蒙古大学图书馆

MINN: learning the Dynamics of Differential-Algebraic Equations and Application to Battery Modeling

ieee TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 2024年第12期46卷 11331-11344页

作者： Huang, Yicun Zou, Changfu Li, Yang Wik, Torsten Chalmers Univ Technol Dept Elect Engn S-41296 Gothenburg Sweden

The concept of integrating physics-based and data-driven approaches has become popular for modeling sustainable energy systems. However, the existing literature mainly focuses on the data-driven surrogates generated to replace physics-based models. These models often trade accuracy for speed but lack the generalizability, adaptability, and interpretability inherent in physics-based models, which are often indispensable in modeling real-world dynamic systems for optimization and control purposes. We propose a novel machine learning architecture, termed model-integrated neural networks (MINN), that can learn the physics-based dynamics of general autonomous or non-autonomous systems consisting of partial differential-algebraic equations (PDAEs). The obtained architecture systematically solves an unsettled research problem in control-oriented modeling, i.e., how to obtain optimally simplified models that are physically insightful, numerically accurate, and computationally tractable simultaneously. We apply the proposed neural network architecture to model the electrochemical dynamics of lithium-ion batteries and show that MINN is extremely data-efficient to train while being sufficiently generalizable to previously unseen input data, owing to its underlying physical invariants. The MINN battery model has an accuracy comparable to the first principle-based model in predicting both the system outputs and any locally distributed electrochemical behaviors but achieves two orders of magnitude reduction in the solution time.

关键词： Lithium-ion batteries battery management systems battery modeling model simplification physics-informed machine learning model-integrated neural networks

来源：评论

学校读者我要写书评

暂无评论

Event-Triggered Sampled-data Iterative learning control for Nonlinear Switched Multi-Agent systems 13

Event-Triggered Sampled-Data Iterative Learning Control for ...

引用

13th ieee data driven control and learning systems conference, DDCLS 2024

作者： Ma, Xin Dai, Xisheng Wang, Guoyong School of Automation Guangxi University of Science and Technology Liuzhou545000 China Liuzhou High School Liuzhou545000 China

ISBN: (纸本)9798350361674

This paper investigates a class of consensus tracking problems in switched multi-agent systems(SMASs). For a virtual leader of switched multi-agent systems, preset an expected virtual leader tracking trajectory. Firstly, this paper designs a class of output tracking information that includes adjacent agents and sampled-data iterative learning control. Then, this paper designs a sampling-based event-triggered iterative learning control(ETILC) law to control the tracking behavior of agents. Reuse a Lyapunov-like method to determine the event-triggered conditions for system stability. Finally, for nonlinear switched multi-agent systems in various switching subsystems, the convergence conditions of systems can be discussed based on triggering and untriggering in different subsystems, using the Bellman-Gronwall inequality and compression mapping method. Verify the effectiveness and feasibility of the proposed algorithm through an instance of multi-agent systems that includes four switching modes. © 2024 ieee.

关键词： Lyapunov methods

来源：评论

学校读者我要写书评

暂无评论

A Variational Bayesian Dictionary learning for Process Monitoring 12

A Variational Bayesian Dictionary Learning for Process Monit...

引用

ieee 12th data driven control and learning systems conference (DDCLS)

作者： Zhang, Qi Xie, Lei Xu, Weihua Su, Hongye Zhejiang Univ State Key Lab Ind Control Technol Hangzhou 310027 Peoples R China

ISBN: (纸本)9798350321050

Industrial process data acquisition is inevitably disturbed by noise, and data contamination makes process data carry too much redundant information, which will greatly limit the interpretation capability of data-driven modeling approaches. Given that sparse representations can effectively handle noise, a process monitoring method for variational Bayesian dictionary learning (VBDL) is developed in this work. Conventional dictionary learning requires a priori knowledge of the assumed noise variance and sparsity levels, which is not available in real industrial processes. The derived VBDL is built on a Bernoulli distribution with the beta distribution as the conjugate prior, the number of dictionary atoms and their relative importance can be inferred nonparametrically with the iterative update of the variational inference. Since the collected data exhibits temporal correlation, the large noise interference makes dynamic analysis infeasible. A low-rank vector autoregression is developed to dynamically analyze the reconstructed samples, thereby improving the robustness of the model to noise. To illustrate the feasibility and efficacy, the proposed algorithm is verified by a numerical simulation and the CSTR simulation.

关键词： Beta-Bernoulli process dictionary learning process monitoring probabilistic model variational Bayesian

来源：评论

学校读者我要写书评

暂无评论

Signal Reconstruction of Output Attacks on Autonomous Vehicles 12

Signal Reconstruction of Output Attacks on Autonomous Vehicl...

引用

ieee 12th data driven control and learning systems conference (DDCLS)

作者： Chen, Li Guo, Shenghui Suzhou Univ Sci & Technol Sch Elect & Informat Engn Suzhou 215009 Peoples R China

ISBN: (纸本)9798350321050

This paper mainly studies the problem of vehicle attack signal estimation when autonomous vehicles contain network attacks and multiple autonomous vehicles communicate with each other. This paper first selects the appropriate vehicle dynamics model and vehicle lane-keeping model to model the vehicle. Then, considering the network attack of the vehicle and the mutual influence of multiple autonomous vehicles, the subsystem model is established, and the subsystem is converted into the form of the whole system. Then, the intermediate variable observer is designed for the whole system, and the output attack signal is reconstructed using the system state estimated by the observer. Finally, an example of two autonomous vehicles communicating with each other is used for simulation analysis to verify the feasibility of the proposed method.

关键词： Cyber-physical systems Autonomous vehicles State estimation Intermediate variable observer Attack reconstruction

来源：评论

学校读者我要写书评

暂无评论

Stability and Stabilization of data-driven Discrete-time Linear Impulsive systems 13

Stability and Stabilization of Data-driven Discrete-time Lin...

引用

13th ieee data driven control and learning systems conference, DDCLS 2024

作者： Zhou, Dongpeng Chen, Wu-Hua Lu, Xiaomei School of Electrical Engineering Guangxi University Guangxi Nanning530004 China School of Mathematics and Information Science Guangxi University Guangxi Nanning530004 China

ISBN: (纸本)9798350361674

This paper studies the problems of stability and stabilization for data-driven linear discrete-time impulsive systems (LDTISs). Firstly, open-loop and closed-loop data-based system representations of LDTISs are given by utilizing persistently exciting data. While the open-loop data-based system representation is an identification-like result, the closed-loop case bypasses the system identification, which can be used for directly data-driven control law design. Secondly, by utilizing time-dependent Lyapunov function technique, stability criteria for data-based closed-loop system are derived, and a data-driven state feedback controller are designed. Such criteria are independent of system matrices and can be formulated into a set of linear matrix inequalities plus a class of linear matrix equations, which can be solved efficiently by CVX toolbox. Finally, an example is presented to verify the efficiency of the established data-driven control law. © 2024 ieee.

关键词： Lyapunov functions

来源：评论

学校读者我要写书评

暂无评论

Improved quantized predictive iterative learning control for systems with variable interval lengths and data dropouts 13

Improved quantized predictive iterative learning control for...

引用

13th ieee data driven control and learning systems conference, DDCLS 2024

作者： Shao, Zhen Xue, Songyi Wang, Ronghao Chen, Yu Duan, Zhaoxia Kong, Fanrong College of Artificial Intelligence and Automation Hohai University Nanjing211100 China College of National Defense Engineering Army Engineering University of Pla Nanjing210007 China

ISBN: (纸本)9798350361674

In this paper, the problem of complete tracking control is investigated for single-input single-output unknown nonlinear discrete systems with variable interval lengths, data quantization and data dropouts. First, a quantized model-free adaptive predictive iterative learning control scheme is proposed based on the dynamic linearization model of the system. The scheme employs the encoding-decoding mechanism and an intermittent update principle to eliminate the inherent errors caused by data quantization and data dropouts, and designs compensation and prediction mechanisms to deal with the missing data problem caused by variable interval lengths. Then, it is demonstrated that convergence of the tracking control error can be achieved by utilizing only a small amount of input and output data. Finally, the effectiveness of the proposed method is verified by simulation. © 2024 ieee.

关键词： Contrastive learning

来源：评论

学校读者我要写书评

暂无评论

Efficient Replay Deep Meta-Reinforcement learning for Active Fault-Tolerant control of Solid Oxide Fuel Cell systems Considering Multivariable Coordination

引用

ieee TRANSACTIONS ON TRANSPORTATION ELECTRIFICATION 2025年第1期11卷 4803-4817页

作者： Li, Jiawen Zhou, Tao Shanghai Univ Elect Power Sch Elect & Informat Engn Shanghai 201306 Peoples R China Shanghai Jiao Tong Univ Sch Elect Informat & Elect Engn Dept Elect Engn Shanghai 200240 Peoples R China

A data-driven integrated active fault-tolerant control (IAFT) strategy for controlling the solid oxide fuel cell (SOFC) output voltage is proposed, which maintains satisfactory dynamic performance and eliminates constraint violations in the event of system failure. In addition, this article introduces an efficient replay deep meta-deterministic policy gradient (ER-DMDPG) for IAFTs, which combines priority experience replay and meta-learning techniques to improve the robustness and multitask cooperative learning capability of the IAFTs. The algorithm combines the controllers of the fuel reformer and direct current-direct current (dc-dc) converter into a single independent agent, which is trained by a cooperative meta-learner and a base learner to achieve multiobjective optimal active fault-tolerant control (FTC). It is experimentally demonstrated that the proposed method can maintain better dynamic performance and prevent constraint violations of fuel utilization across a wide range of working conditions.

关键词： Fault tolerant systems Fault tolerance Hydrogen Voltage control Fuels Fuel cells Temperature control Prediction algorithms Heuristic algorithms Genetic algorithms Active fault-tolerant integrated control constraint violations deep meta-reinforcement learning (DMRL) fuel utilization solid oxide fuel cell (SOFC)

来源：评论

学校读者我要写书评

暂无评论

Reinforcement-learning-aided adaptive control for autonomous driving with combined lateral and longitudinal dynamics 12

Reinforcement-learning-aided adaptive control for autonomous...

引用

ieee 12th data driven control and learning systems conference (DDCLS)

作者： Wang, Yongshuai Zheng, Chen Sun, Mingwei Chen, Zengqiang Sun, Qinglin Nankai Univ Coll Artificial Intelligence Tianjin 300350 Peoples R China Beijing Inst Astronaut Syst Engn Beijing 100076 Peoples R China

ISBN: (纸本)9798350321050

This paper presents a deep reinforcement learning-aided controller for a 3-DOF autonomous vehicle with combined lateral and longitudinal dynamics. In this scheme, the active disturbance rejection control (ADRC) gives full play to its advantages of being model-free and being able to estimate and compensate for internal uncertainties and external disturbances in real-time, and deep deterministic policy gradient (DDPG) fully considers safety, comfort, economy, and combines driving demand with state, action, reward to achieve real-time adaptive adjustment of control parameters. Thus, the adaptive controller can better deal with uncertainties from modeling, parameters, and driving environment, and self-learning and adaptation ability is obtained simultaneously. Moreover, simulation results illustrate that the adaptive controller performs satisfactorily for different driving operations and environments due to the online tuning and optimization of control parameters.

关键词： autonomous vehicle deep reinforcement learning active disturbance rejection control adaptive tuning

来源：评论

学校读者我要写书评

暂无评论

Spectral Normalized Neural Networks Funnel control for Servo System with Unknown Dynamics 12

Spectral Normalized Neural Networks Funnel Control for Servo...

引用

ieee 12th data driven control and learning systems conference (DDCLS)

作者： Zhang, Chao Ren, Xuemei Han, Ning Zheng, Dongdong Beijing Inst Technol Sch Automat Beijing 100081 Peoples R China

ISBN: (纸本)9798350321050

This paper proposes a novel spectral normalized neural networks funnel control approach for servo system with unknown dynamics. The approach introduces spectral normalization technology into the funnel controller design to address the unknown dynamics. Spectral normalization techniques can restrict the spectral norm of the weight matrices of the neural networks, leading to more stable and robust networks. The spectral normalized neural network exhibits strong generalization ability and can adapt to offline learning strategies, which significantly reduce the system's computation cost. Moreover, based on the funnel control architecture, the system output is constrained to remain within an acceptable boundary, optimizing transient performance and guaranteeing satisfactory control performance. All signals of the closed-loop system are bounded based on Lyapunov stability analysis. Finally, simulation results demonstrate that this approach provides commendable tracking performance and superior generalization capabilities.

关键词： Servo system funnel control spectral normalized neural networks unknown dynamics

来源：评论

学校读者我要写书评

暂无评论

Optimal data-driven Difference-Inversion-Based Iterative control: High-Speed Nanopositioning Tracking Example

引用

ieee TRANSACTIONS ON control systems TECHNOLOGY 2023年第1期31卷 144-154页

作者： Zhang, Zezhou Zou, Qingze Rutgers State Univ Dept Mech & Aerosp Engn Piscataway NJ 08854 USA

In this article, an optimal data-driven difference-inversion-based iterative control (ODDD-IIC) method is proposed for high-speed precision tracking in the presence of dynamics changes and random disturbances. Iterative learning control (ILC) has been shown to be advantageous over feedback and feedforward control for repetitive operations. Challenges, however, still exist to achieve high accuracy and fast convergence in ILCs as the bandwidth, i.e., the frequency range for guaranteed convergence, can be limited by adverse effects of modeling error and random disturbances. The aim of the proposed method is to compensate for these adverse effects through a data-driven approach without a modeling process. A frequency- and iteration-dependent iteration gain is introduced in the control law to enhance both the tracking performance and the robustness. The technique is illustrated in an output tracking experiment on a piezoelectric actuator, with comparison to two existing ILC methods.

关键词： Convergence Robustness System dynamics Nanopositioning data models Uncertainty Aerodynamics data-driven iterative learning control (ILC) nanopositioning control system inversion

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：