检索结果-内蒙古大学图书馆

EXPLOITING EQUIVARIANCE IN THE DESIGN OF TRACKING controlLERS FOR EULER-POINCARE SYSTEMS ON MATRIX LIE GROUPS

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Hampsey, Matthew Banavar, Ravi van Goor, Pieter Mahony, Robert Systems Theory and Robotics Group Australian National University ACT2601 Australia Interdisciplinary Programme in Systems and Control Engineering Indian Institute of Technology Bombay India

The trajectory tracking problem is a fundamental control task in the study of mechanical systems. A key construction in tracking control is the error or difference between an actual and desired trajectory. This construction also lies at the heart of observer design and recent advances in the study of equivariant systems have provided a template for global error construction that exploits the symmetry structure of a group action if such a structure exists. Hamiltonian systems are posed on the cotangent bundle of configuration space of a mechanical system and symmetries for the full cotangent bundle are not commonly used in geometric control theory. In this paper, we propose a group structure on the cotangent bundle of a Lie-group and leverage this to define momentum and configuration errors for trajectory tracking drawing on recent work on equivariant observer design. We show that this error definition leads to error dynamics that are themselves "Euler-Poincare like" and use these to derive simple, almost global trajectory tracking control for fully-actuated Euler-Poincare systems on a Lie-group state space. Copyright © 2024, The Authors. All rights reserved.

关键词： Lie groups

A High-Precision Edge Detection Algorithm Based on Improved Sobel Operator-Assisted HED 14

学校读者我要写书评

暂无评论

A High-Precision Edge Detection Algorithm Based on Improved ...

14th IEEE International Conference on Cyber Technology in Automation, control, and Intelligent Systems, CYBER 2024

作者： Yuan, Yeborui Chen, Wenyuan Tang, Jie Yang, Jia School of Electrical Engineering and Automation Tianjin University of Technology Tianjin300384 China Shenyang Institute of Automation Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences Shenyang State Key Laboratory of Robotics 110016 China University of Chinese Academy of Sciences Beijing100049 China School of Electrical Engineering and Automation Tianjin University of Technology Tianjin Key Laboratory for Control Theory and Applications in Complicated Systems Tianjin300384 China

ISBN: (纸本)9798331506056

In response to the problem of poor detection performance of the traditional Sobel operator in edge detection, a high-precision edge detection algorithm based on Sobel operator-assisted Holistically-nested Edge Detection (HED) network was proposed, which is named ISAHED here. The algorithm is designed as follows. Firstly, an improved Sobel operator based on the traditional Sobel operator was proposed through adding gradient direction. Then, the image output by the improved Sobel operator is taken as the ground-truth image and as the end-to-end comparison object in a HED-like network. Finally, the ISAHED algorithm outputs edge images with better quality through iterative learning of the HED-like neural network. The experimental results prove that the proposed ISAHED algorithm has better edge detection performance by visually observing the final edge images and further evaluating the values of FOM. This work provides an important reference for the further development of the edge detection technology. © 2024 IEEE.

关键词： Contrastive Learning

LipKernel: Lipschitz-Bounded Convolutional Neural Networks via Dissipative Layers

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Pauli, Patricia Wang, Ruigang Manchester, Ian Allgöwer, Frank Institute for Systems Theory and Automatic Control University of Stuttgart Stuttgart70550 Germany Australian Centre for Robotics School of Aerospace Mechanical and Mechatronic Engineering The University of Sydney Australia

We propose a novel layer-wise parameterization for convolutional neural networks (CNNs) that includes built-in robustness guarantees by enforcing a prescribed Lipschitz bound. Each layer in our parameterization is designed to satisfy a linear matrix inequality (LMI), which in turn implies dissipativity with respect to a specific supply rate. Collectively, these layer-wise LMIs ensure Lipschitz boundedness for the input-output mapping of the neural network, yielding a more expressive parameterization than through spectral bounds or orthogonal layers. Our new method LipKernel directly parameterizes dissipative convolution kernels using a 2-D Roesser-type state space model. This means that the convolutional layers are given in standard form after training and can be evaluated without computational overhead. In numerical experiments, we show that the run-time using our method is orders of magnitude faster than state-of-the-art Lipschitz-bounded networks that parameterize convolutions in the Fourier domain, making our approach particularly attractive for improving robustness of learning-based real-time perception or control in robotics, autonomous vehicles, or automation systems. We focus on CNNs, and in contrast to previous works, our approach accommodates a wide variety of layers typically used in CNNs, including 1-D and 2-D convolutional layers, maximum and average pooling layers, as well as strided and dilated convolutions and zero padding. However, our approach naturally extends beyond CNNs as we can incorporate any layer that is incrementally dissipative. Copyright © 2024, The Authors. All rights reserved.

关键词： Parameterization

Lipschitz-Bounded 1D Convolutional Neural Networks using the Cayley Transform and the controllability Gramian

学校读者我要写书评

暂无评论

Lipschitz-Bounded 1D Convolutional Neural Networks using the...

IEEE Conference on Decision and control

作者： Patricia Pauli Ruigang Wang Ian R. Manchester Frank Allgöwer Institute for Systems Theory and Automatic Control University of Stuttgart Stuttgart Germany Australian Centre for Robotics School of Aerospace Mechanical and Mechatronic Engineering The University of Sydney Australia

We establish a layer-wise parameterization for 1D convolutional neural networks (CNNs) with built-in end-to-end robustness guarantees. In doing so, we use the Lipschitz constant of the input-output mapping characterized by a CNN as a robustness measure. We base our parameterization on the Cayley transform that parameterizes orthogonal matrices and the controllability Gramian of the state space representation of the convolutional layers. The proposed parameterization by design fulfills linear matrix inequalities that are sufficient for Lipschitz continuity of the CNN, which further enables unconstrained training of Lipschitz-bounded 1D CNNs. Finally, we train Lipschitz-bounded 1D CNNs for the classification of heart arrythmia data and show their improved robustness.

关键词：

Lipschitz-bounded 1D convolutional neural networks using the Cayley transform and the controllability Gramian

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Pauli, Patricia Wang, Ruigang Manchester, Ian R. Allgöwer, Frank The Institute for Systems Theory and Automatic Control University of Stuttgart Stuttgart70569 Germany The Australian Centre for Robotics School of Aerospace Mechanical and Mechatronic Engineering The University of Sydney Australia

关键词： State space methods

Reprojection Methods for Koopman-Based Modelling and Prediction

学校读者我要写书评

暂无评论

Reprojection Methods for Koopman-Based Modelling and Predict...

IEEE Conference on Decision and control

作者： Pieter Van Goor Robert Mahony Manuel Schaller Karl Worthmann Systems Theory and Robotics Group School of Engineering Australian National University Canberra Australia Optimization-based Control Group Institute of Mathematics Technische Universitat Ilmenau Ilmenau Germany

Extended Dynamic Mode Decomposition (eDMD) is a powerful tool to generate data-driven surrogate models for the prediction and control of nonlinear dynamical systems in the Koopman framework. In eDMD a compression of the lifted system dynamics on the space spanned by finitely many observables is computed, in which the original space is embedded as a low-dimensional manifold. While this manifold is invariant for the infinite-dimensional Koopman operator, this invariance is typically not preserved for its eDMD-based approximation. Hence, an additional (re-)projection step is often tacitly incorporated to improve the prediction capability. We propose a novel framework for consistent reprojectors respecting the underlying manifold structure. Further, we present a new geometric reprojector based on maximum-likelihood arguments, which significantly enhances the approximation accuracy and preserves known finite-data error bounds.

关键词：

Reprojection methods for Koopman-based modelling and prediction

学校读者我要写书评

暂无评论

arXiv 2023年

作者： van Goor, Pieter Mahony, Robert Schaller, Manuel Worthmann, Karl Systems Theory and Robotics Group School of Engineering Australian National University Canberra Australia Optimization-based Control Group Institute of Mathematics Technische Universität Ilmenau Germany

关键词： Dynamic mode decomposition

SLAM-Based Joint Calibration of Differential RSS Sensor Array and Source Localization 49

学校读者我要写书评

暂无评论

SLAM-Based Joint Calibration of Differential RSS Sensor Arra...

49th Annual Conference of the IEEE Industrial Electronics Society, IECON 2023

作者： Fu, Linya Qiao, Xu Huang, Shoudong Mao, Guoqiang Lin, Zhiyun Li, Youfu Kong, He Robotics Institute Uni. of Technology Sydney Sydney Australia Xidian University State Key Laboratory of Integrated Services Networks China City University of Hong Kong Department of Mechanical Engineering China Southern Uni. of Sci. and Tech. Shenzhen Key Laboratory of Control Theory and Intelligent Systems Shenzhen China Southern Uni. of Sci. and Tech. Guangdong Prov. Key Laboratory of Human-Augmentation and Rehabilitation Robotics in Universities Shenzhen China

ISBN: (纸本)9798350331820

Sensor arrays generating differential received signal strength (DRSS) measurements have found many applications in robotics. However, accurate calibration of these sensor arrays remains a challenge. Most existing methods are impractical in that they assume to know signal source positions or certain parameters (i.e., path loss exponent), and try to estimate the others. In this paper, we adopt graph simultaneous localization and mapping (SLAM) as a general framework for jointly estimating the source positions and parameters of the DRSS sensor array. Our contributions are twofold. On the one hand, by using a Fisher information matrix approach, we conduct a systematic observability analysis of the corresponding SLAM setup for the calibration problem. On the other hand, we propose an effective procedure to select the initial value which is fed to Levenberg-Marquardt iterations for further improving optimization accuracy and convergence. Extensive simulation and hardware experiments show that the proposed method renders high-quality calibration results. All the codes and data are publicly available at https://***/SUSTech2022/DRSS-sensor-array-calibration. © 2023 IEEE.

关键词： Calibration

Dynamic modelling and PFL-based trajectory tracking control for underactuated cable-driven truss-like manipulator

学校读者我要写书评

暂无评论

Journal of Central South University 2021年第10期28卷 3127-3146页

作者： DING Shu-chen PENG Li QIAO Shang-ling LIU Rong-qiang JOSEPHAT Bundi School of IoT Engineering Jiangnan UniversityWuxi 214122China State Key Laboratory of Robotics and System Harbin Institute of TechnologyHarbin 150001China Center for Control Theory and Guidance Technology Harbin Institute of TechnologyHarbin 150001China

In recent years,an innovative underactuated robot was developed,named as underactuated cable-driven trusslike manipulator(UCTM),to be suitable in aerospace ***,there has been strong consensus that the stabilization of planar underactuated manipulators without gravity is a great challenge since the system includes a second order nonholonomic constraint and most classical control methods are not suitable for this kind of ***,the complexity of the truss-like structure results in tremendous difficulty of computational complicacy and high nonlinearity during dynamic modelling in addition to controller *** is paramount to solve these difficulties for UCTM's future *** solve the above difficulties,this paper presents a dynamic modelling method for UCTM and a trajectory tracking control method based on partial feedback linearization(PFL)that fulfills the control goal of moving UCTM from its original position to a desired position by tracking a given trajectory of the joint *** achieve this,a model equivalent method is proposed to make UCTM equivalent with a three-link manipulator in the sense of dynamic *** the Lagrangian equation combined with complex vector method is proposed in the dynamic modelling process of UCTM,which simplifies the derivation *** on the established dynamic model,a coordinate transformation method is proposed to transform the control force matrix into the conventional form of an underactuated system,so that the control force can be separated from the unactuated *** PFL method in combination with the LQR control method is then proposed to realize the targets that the joint angles can track given desired *** experiments are conducted to verify the correctness and effectiveness of the proposed methods.

关键词： underactuated robot trajectory tracking control partial feedback linearization non-linear control