检索结果-内蒙古大学图书馆

10th International Conference on Learning Representations, ICLR 2022

作者： Han, Minghao Euler-Rolle, Jacob Katzschmann, Robert K. Department of Control Science and Engineering Harbin Institute of Technology China Soft Robotics Lab ETH Zurich Switzerland

The Koopman operator theory linearly describes nonlinear dynamical systems in a high-dimensional functional space and it allows to apply linear control methods to highly nonlinear systems. However, the Koopman operator does not account for any uncertainty in dynamical systems, causing it to perform poorly in real-world applications. Therefore, we propose a deep stochastic Koopman operator (DeSKO) model in a robust learning control framework to guarantee stability of nonlinear stochastic systems. The DeSKO model captures a dynamical system's uncertainty by inferring a distribution of observables. We use the inferred distribution to design a robust, stabilizing closed-loop controller for a dynamical system. Modeling and control experiments on several advanced control benchmarks show that our framework is more robust and scalable than state-of-the-art deep Koopman operators and reinforcement learning methods. Tested control benchmarks include a soft robotic arm, a legged robot, and a biological gene regulatory network. We also demonstrate that this robust control method resists previously unseen uncertainties, such as external disturbances, with a magnitude of up to five times the maximum control input. Our approach opens up new possibilities in learning control for high-dimensional nonlinear systems while robustly managing internal or external uncertainty. © 2022 ICLR 2022 - 10th International Conference on Learning Representationss. All rights reserved.

关键词： Dynamical systems

来源：评论

学校读者我要写书评

暂无评论

An Obstacle Avoidance control Strategy for DELTA Parallel Manipulator via Improved Genetic Algorithm 37

An Obstacle Avoidance Control Strategy for DELTA Parallel Ma...

引用

37th Youth Academic Annual Conference of Chinese Association of Automation, YAC 2022

作者： Zheng, Hao Xu, Yihang Li, Yuantao Ran, Guangtao Liu, Jian Southeast University School of Automation Nanjing210018 China Harbin Institute of Technology Department of Control Science and Engineering Harbin150001 China

ISBN: (纸本)9781665465366

Aiming at the requirement of the DELTA parallel manipulator for performing grasping operations, a method for obstacle avoidance based on a genetic algorithm is proposed. Firstly, the classical DELTA parallel manipulator is analyzed mathematically, and its model is simplified. Then, the motion space and inverse kinematics solutions of the DELTA parallel manipulator are solved by using the spatial geometry method. Secondly, a cylindrical envelope surface is added to the manipulator's end for distance calculation. This envelope surface ensures that obstacles do not collide with the manipulator during obstacle avoidance. Thirdly, an improved genetic algorithm is proposed for obstacle avoidance. Finally, the simulation results are given to verify the effectiveness and reliability of the proposed algorithm. © 2022 IEEE.

关键词： Genetic algorithms

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Learning Based control for Uncertain Robotic Manipulator Trajectory Tracking

Reinforcement Learning Based Control for Uncertain Robotic M...

引用

2022 Chinese Automation Congress, CAC 2022

作者： Liu, Aohua Zhang, Bo Chen, Weiliang Luo, Yiang Fang, Shuxian Zhang, Ouyang Liu, Zhuang Wang, Zhenhuan Liu, Jianxing Harbin Institute of Technology Department of Control Science and Engineering Harbin150001 China State Grid Heilongjiang Electric Power Company Limited The Suihua Power Supply Company Suihua China Harbin Institute of Technology Space Control and Inertial Technology Research Center Harbin150001 China

ISBN: (纸本)9781665465335

This paper investigates the trajectory tracking control method for a robotic manipulator with uncertainties. A compound controller combining a traditional control law with deep reinforcement learning is developed to improve tracking accuracy and adaptability. The model-based control method is able to increase sampling efficiency for learning control strategy. The introduction of deep reinforcement learning based on soft actor-critic structure and Lyapunov function constrain enables the system to compensate unknown uncertainties and remain stable. Eventually, a 3-DOF manipulator is used to show the effectiveness of the proposed controller. Comparative simulation results demonstrate that the compound controller acquires higher tracking accuracy than the pure model-based control method. © 2022 IEEE.

关键词： controllers

来源：评论

学校读者我要写书评

暂无评论

Robust control for DC-DC Buck Converters Based on Fully Actuated System theory

Robust Control for DC-DC Buck Converters Based on Fully Actu...

引用

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

作者： Fangyuan Li Guangxin Liu Xiaoning Shen Cheng Li Xinpo Lin Yabin Gao Jianxing Liu Department of Control Science and Engineering Harbin Institute of Technology Harbin P. R. China

ISBN: (数字)9798350373691

ISBN: (纸本)9798350373707

This paper presents a novel robust approach for buck converters voltage tracking by utilizing the high-order fully actuated (HOFA) theory. Initially, the HOFA model of the buck converter is derived, followed by an analysis of disturbances encountered in real-world scenarios. In order to mitigate the adverse consequences arising from imprecise or abrupt alterations in circuit parameters, a robust tracking controller is devised to enhance the overall robustness of the system. Ultimately, simulations are performed to compare the proposed controller with the dual-loop PI controller. The results indicate that the proposed controller exhibits a quicker dynamic response and greater robustness compared to the dual-loop PI controller. Additionally, the process of selecting parameters of this method is simpler and more convenient.

关键词： Robust control Resistance Buck converters PI control Upper bound System performance Process control

来源：评论

学校读者我要写书评

暂无评论

Investigating Deep Watermark Security: An Adversarial Transferability Perspective

arXiv

引用

arXiv 2024年

作者： Qi, Biqing Gao, Junqi Luo, Yiang Liu, Jianxing Wu, Ligang Zhou, Bowen Department of Control Science and Engineering Harbin Institute of Technology Harbin China Department of Electronic Engineering Tsinghua University Bejing China School of Mathematics Harbin Institute of Technology Harbin China

The rise of generative neural networks has triggered an increased demand for intellectual property (IP) protection in generated content. Deep watermarking techniques, recognized for their flexibility in IP protection, have garnered significant attention. However, the surge in adversarial transferable attacks poses unprecedented challenges to the security of deep watermarking techniques—an area currently lacking systematic investigation. This study fills this gap by introducing two effective transferable attackers to assess the vulnerability of deep watermarks against erasure and tampering risks. Specifically, we initially define the concept of local sample density, utilizing it to deduce theorems on the consistency of model outputs. Upon discovering that perturbing samples towards high sample density regions (HSDR) of the target class enhances targeted adversarial transferability, we propose the Easy Sample Selection (ESS) mechanism and the Easy Sample Matching Attack (ESMA) method. Additionally, we propose the Bottleneck Enhanced Mixup (BEM) that integrates information bottleneck theory to reduce the generator’s dependence on irrelevant noise. Experiments show a significant enhancement in the success rate of targeted transfer attacks for both ESMA and BEM-ESMA methods. We further conduct a comprehensive evaluation using ESMA and BEM-ESMA as measurements, considering model architecture and watermark encoding length, and achieve some impressive findings. Copyright © 2024, The Authors. All rights reserved.

关键词： Watermarking

来源：评论

学校读者我要写书评

暂无评论

Research on multi-dimensional heterogeneous compound control method of anti-access aircraft with prescribed performance

Research on multi-dimensional heterogeneous compound control...

引用

第43届中国控制会议

作者： Yuemeng Ma Ningyuan Wang Ming Liu Zhen Xiao Ding Yang Jingying Cao Science and Technology on Space Physics Laboratory Department of Control Science and Engineering Harbin Institute of Technology

ISBN: (数字)9789887581581

ISBN: (纸本)9798350366907

For attack and defense confrontation scenarios in near-space,to enhance the maneuverability of the anti-access aircraft in low dynamic pressure environment,a prescribedperformance-based multidimensional heterogeneous compound control system is proposed in this paper to meet the precision guidance requirements of fast and high-precision response to overload commands by the synergistic cooperation of direct force and aerodynamic *** time and steady state accuracy of the system are designed based on the prescribed performance control *** surface controllers is designed for aerofin to observe and compensate for complex *** direct forces,a sliding mode controller is designed and applied to achieve rapid tracking to *** layers and auxiliary system are introduced to reduce the impact of actuator physical constraints on control *** Lyapunov stability proof of compound control system is *** results demonstrate that the multidimensional heterogeneous compound control system proposed can not only make the response process of AOA meet the prescribed-performance requirement but also avoid the "tug-of-war" phenomenon of actuators.

关键词： anti-access aircraft multi-dimensional heterogeneous compound control lateral jets and aerodynamics control prescribed performance control nonlinear extended state observer(NESO)

来源：评论

学校读者我要写书评

暂无评论

Combined Spacecraft Attitude Tracking control with Prescribed Performance via A Fully Actuated System Approach

Combined Spacecraft Attitude Tracking Control with Prescribe...

引用

第42届中国控制会议

作者： Guang-Quan Duan Guo-Ping Liu Yan-Ming Fu Center for Control Theory and Guidance Technology Harbin Institute of Technology Department of Electronic and Electrical Engineering Southern University of Science and Technology

In this paper,a fully-actuated system prescribed performance controller is proposed to achieve attitude tracking control of the combined *** order to make the combined spacecraft satisfy the required steadystate and transient performance in the attitude tracking control process,a fully actuated system controller with prescribed performance is *** attitude dynamics model of the combined spacecraft is transformed into a tracking error ***,the controller of the fully actuated system is designed according to the error function to realize the attitude tracking control of the combined *** function is introduced to prove the stability of the *** the presented way of designing fully actuated system prescribed performance controller,the numerical simulation results verify the effectiveness of the proposed method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Digital Twin Model Design of Three-phase Inverter for Reliability Prediction 8

Digital Twin Model Design of Three-phase Inverter for Reliab...

引用

8th International Conference on Reliability of Electrical Products and Electrical Contacts, ICREPEC 2022

作者： Chen, Cen Sun, Chengzhi Zheng, Wei Zhang, Yuchen Xu, Yi Ye, Xuerong Zhai, Guofu Department of Electric Engineering and Automation Harbin Institute of Technology Harbin150001 China Beijing Aerospace Automatic Control Research Institute Beijing100089 China

ISBN: (纸本)9789881532374

Three-phase inverter is widely used in DC-AC power supply, motor drive, photovoltaic inverter, etc., whose reliability and stability are fundamental of efficient operation of various electrical systems. Digital twin technology can realize the life cycle performance analysis of the product by ingeniously combining data derived from multi-source and multi-level. Aiming at the reliability prediction, Saber, ANSYS and other software are used as carriers, and a 270V DC three-phase inverter is carried out to study the method of establishing the digital prototype model by integrating circuit models, behavior models and physical models. On this basis, taking DC/DC modules, diodes, MOSFETs, operational amplifiers, drivers as examples, the equivalent injection method of physics of failure models integrating performance degradation and functional failure is studied. Finally, a test method for products is researched to verify the validity and accuracy of the three-phase inverter digital twin model. © ICREPEC *** rights reserved

关键词： Digital devices

来源：评论

学校读者我要写书评

暂无评论

Sparse Signal Representation-Based Absorption Coefficients Extraction for Photoacoustic Tomography

SSRN

引用

SSRN 2023年

作者： Liu, Yang Ma, Yiming Qin, Zezheng Sun, Mingjian Department of Control Science and Engineering Harbin Institute of Technology Heilongjiang Harbin150001 China Department of Control Science and Engineering Harbin Institute of Technology Shandong Weihai264200 China

In quantitative photoacoustic tomography (qPAT), the raw photoacoustic images only indirectly describe structural and physiological information on biological tissues. Therefore, for specific biological tissue models, it is necessary to study a high-precision absorption coefficient (AC) extraction algorithm to convert the photoacoustic image into the absorption coefficient distribution map of the absorber by eliminating the influence of non-uniform light fluence (LF). The absorption coefficient distribution is directly related to the spatial distribution of the absorbing chromophore in the tissue. In this study, a new quantitative photoacoustic imaging algorithm based on sparse signal representation is introduced to extract AC of multi-absorbers with higher accuracy. We take advantage of the difference between the spatial distribution characteristics of AC and LF in the medium to obtain dictionary libraries that can sparsely represent AC and LF, respectively, and combine them into a joint dictionary. Then, based on sparse representation theory, the photoacoustic image is decomposed, and the AC of the imaging object is further extracted from the photoacoustic image. Simulation results show that our method achieves better AC image accuracy and feature visibility in complex cases with multi-absorbers. Our method is expected to be applied in the future to the quantitative analysis of photoacoustic imaging, which requires accurate estimation of the absorption distribution of biological tissues containing multiple absorbers, such as vascular system and tumor detection. © 2023, The Authors. All rights reserved.

关键词： Spatial distribution

来源：评论

学校读者我要写书评

暂无评论

Certificated Actor-Critic: Hierarchical Reinforcement Learning with control Barrier Functions for Safe Navigation

arXiv

引用

arXiv 2025年

作者： Xie, Junjun Zhao, Shuhao Hu, Liang Gao, Huijun Department of Automation School of Mechanical Engineering and Automation Harbin Institute of Technology Shenzhen China Research Institute of Intelligent Control and Systems Harbin Institute of Technology Harbin China

control Barrier Functions (CBFs) have emerged as a prominent approach to designing safe navigation systems of robots. Despite their popularity, current CBF-based methods exhibit some limitations: optimization-based safe control techniques tend to be either myopic or computationally intensive, and they rely on simplified system models;conversely, the learning-based methods suffer from the lack of quantitative indication in terms of navigation performance and safety. In this paper, we present a new model-free reinforcement learning algorithm called Certificated Actor-Critic (CAC), which introduces a hierarchical reinforcement learning framework and well-defined reward functions derived from CBFs. We carry out theoretical analysis and proof of our algorithm, and propose several improvements in algorithm implementation. Our analysis is validated by two simulation experiments, showing the effectiveness of our proposed CAC algorithm. Copyright © 2025, The Authors. All rights reserved.

关键词： Adversarial machine learning

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：