检索结果-内蒙古大学图书馆

International Symposium on Intelligent Systems and Informatics (SISY)

作者： József Bokor Zoltán Szabó HUN-REN Institute for Computer Science and Control SZTAKI Budapest Hungary Research Center of Vehicle Industry (SZE-JKK) Széchenyi István University Gyor Hungary

ISBN: (数字)9798350385601

ISBN: (纸本)9798350385618

Modeling dynamic systems is an important part in analysing and control of various systems arising either in health sciences or in the engineering word. Recent approaches to learn models from data are the so-called kernel-based methods and SVMs. There are strong relations to the theory of reproducing kernel Hilbert space (RKHS), to principal component analysis and canonic correlation analysis known previously from statistics. In recent form their use was extended from statistics to obtain models for dynamic systems. First we summarise the basics for the reproducing kernel based Hilbert space (RKHS) and the support vector machine (SVM) approaches. Following this it will be shown how some frequently used nonlinear models can be obtained by using these concepts. In the last part we discuss the structure estimation problem, i.e. how to determine the (least) number of features (observables) to describe the nonlinear system with a sparse representation.

关键词： Support vector machines Analytical models Sparse approximation Statistical learning Data models Hilbert space Eigenvalues and eigenfunctions Kernel Dynamical systems Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

Estimation of cornering stiffness using ultra-local model and LPV-based observer

引用

IFAC-PapersOnLine 2024年第4期58卷 437-442页

作者： Dániel Fényes Tamás Hegedűs Péter Gáspár HUN-REN Institute for Computer Science and Control (SZTAKI) Kende u. 13-17 H-1111 Budapest Hungary Department for Control of Transportationand Vehicle Systems Budapest University of Technology and Economics Stoczek u 2. H-1111 Budapest Hungary

The paper presents an observer design method for estimating the slips and the cornering stiffness of autonomous road vehicles. As a first step, the dynamical one-wheeled bicycle model is reformulated to express directly the slips of the vehicle. In the second step, a polytopic model-based observer is designed which can deal with the variation of the scheduling parameters (longitudinal velocity, cornering stiffness). In order to eliminate the parameter uncertainty an ultra-local model is applied by which the cornering stiffness is estimated. The effectiveness and the operation of the proposed observer method are demonstrated through a complex scenario in the high-fidelity simulation software, CarMaker.

关键词： ultra-local model observer lateral dynamics cornering stiffness

来源：评论

学校读者我要写书评

暂无评论

State Estimation Design and Tuning for Autonomous Vehicles in the Era of Data-Based Techniques

State Estimation Design and Tuning for Autonomous Vehicles i...

引用

International Symposium on Intelligent Systems and Informatics (SISY)

作者： Máté Fazekas Péter Gáspár Systems and Control Lab HUN-REN Institute for Computer Science and Control Hungarian Research Network (HUN-REN SZTAKI) Budapet Hungary Department of Control for Transportation and Vehicle Systems Budapest University of Technology and Economics (BME) Budapest Hungary

ISBN: (数字)9798350385601

ISBN: (纸本)9798350385618

In the era of autonomous vehicles, state estimation is crucial for planning and control. While various sensors like GNSS, IMU, cameras, and wheel encoders provide essential data, each has limitations. Sensor fusion, especially combining GNSS and IMU, addresses these challenges. Recently, data-based techniques using machine learning tools have emerged to enhance tuning processes. However, if an architecture that has a practical impact is utilized, the training of the neural net results in a complicated task. This paper explores different state estimation architectures in the era of data-based techniques and proposes an algorithm generating reference tuning values. The presented methods are tested with real vehicle measurements.

关键词： Training Global navigation satellite system Neural networks Wheels Sensor fusion Sensors Planning State estimation Tuning Autonomous vehicles

来源：评论

学校读者我要写书评

暂无评论

Nonparametric, Nonasymptotic Confidence Bands with Paley-Wiener Kernels for Band-Limited Functions

arXiv

引用

arXiv 2022年

作者： Csáji, Balázs Csanád Horváth, Bálint SZTAKI: Institute for Computer Science and Control Budapest Hungary Budapest Hungary Budapest Hungary Hungary

The paper introduces a method to construct confidence bands for bounded, band-limited functions based on a finite sample of input-output pairs. The approach is distribution-free w.r.t. the observation noises and only the knowledge of the input distribution is assumed. It is nonparametric, that is, it does not require a parametric model of the regression function and the regions have non-asymptotic guarantees. The algorithm is based on the theory of Paley-Wiener reproducing kernel Hilbert spaces. The paper first studies the fully observable variant, when there are no noises on the observations and only the inputs are random;then it generalizes the ideas to the noisy case using gradient-perturbation methods. Finally, numerical experiments demonstrating both cases are presented. Copyright © 2022, The Authors. All rights reserved.

关键词： Stochastic systems

来源：评论

学校读者我要写书评

暂无评论

Theoretical Evaluation of Asymmetric Shapley Values for Root-Cause Analysis

arXiv

引用

arXiv 2023年

作者： Kelen, Domokos M. Petreczky, Mihály Kersch, Péter Benczúr, András A. HUN-REN SZTAKI Hungarian Research Network Institute for Computer Science and Control Budapest Hungary CNRS University Lille Lille France Ericsson Hungary Budapest Hungary

In this work, we examine Asymmetric Shapley Values (ASV), a variant of the popular SHAP additive local explanation method. ASV proposes a way to improve model explanations incorporating known causal relations between variables, and is also considered as a way to test for unfair discrimination in model predictions. Unexplored in previous literature, relaxing symmetry in Shapley values can have counter-intuitive consequences for model explanation. To better understand the method, we first show how local contributions correspond to global contributions of variance reduction. Using variance, we demonstrate multiple cases where ASV yields counter-intuitive attributions, arguably producing incorrect results for root-cause analysis. Second, we identify generalized additive models (GAM) as a restricted class for which ASV exhibits desirable properties. We support our arguments by proving multiple theoretical results about the method. Finally, we demonstrate the use of asymmetric attributions on multiple real-world datasets, comparing the results with and without restricted model families using gradient boosting and deep learning models. Copyright © 2023, The Authors. All rights reserved.

关键词： Additives

来源：评论

学校读者我要写书评

暂无评论

Terrain Depth Estimation for Improved Inertial Data Prediction in Autonomous Navigation Systems

Terrain Depth Estimation for Improved Inertial Data Predicti...

引用

Automated Vehicle Validation Conference (IAVVC), IEEE International

作者： Norbert Markó Tamás Szirányi Áron Ballagi Machine Perception Research Laboratory Institute for Computer Science and Control (SZTAKI) Budapest Hungary Department of Automation Széchenyi István University Gyὄr Hungary

The prediction of terrain elevation values is a key task when it comes to off-road dynamics and inertial data estimation. A reliable elevation map can help in the estimation of future vehicle states and thus extend the response time window for autonomous navigation and control. We trained a deep learning model that is able to successfully predict top-down terrain depth maps in an off-road setting using a lightweight monocular depth estimation network. The labels were generated using a custom preprocessing algorithm to aid single image depth model training. Unlike other elevation estimation algorithms, our work can predict terrain variation from a higher camera setting without the use of a multi-sensor system. The network is also shown to work outside of the training data domain. The code will be available at https://***/norbertmarko/terrain-depth.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Theoretical Evaluation of Asymmetric Shapley Values for Root-Cause Analysis

Theoretical Evaluation of Asymmetric Shapley Values for Root...

引用

IEEE International Conference on Data Mining (ICDM)

作者： Domokos M. Kelen Mihàly Petreczky Péter Kersch Andràs A. Benczúr HUN-REN SZTAKI Hungarian Research Network Institute for Computer Science and Control Budapest Hungary CNRS University Lille Lille France Ericsson Hungary Budapest Hungary

关键词：

来源：评论

学校读者我要写书评

暂无评论

Enhanced LPV-based lateral control using ultra-local model-based slip angle estimation

Enhanced LPV-based lateral control using ultra-local model-b...

引用

IEEE Conference on Decision and control

作者： Dániel Fényes Tamás Hegedűs Péter Gáspár Hungarian Research Network (HUN-REN) Kende u. 13-17 Institute for Computer Science and Control (SZTAKI) Budapest Hungary Department for Control of Transportation and Vehicle Systems Budapest University of Technology and Economics Stoczek u 2 Budapest Hungary

ISBN: (数字)9798350316339

ISBN: (纸本)9798350316346

The paper presents a combined observer and lateral control design approach for autonomous vehicles. The goal of the observer design is to estimate the front and rear slip angles of the vehicle together with the cornering stiffness. The observer is based on the combination of the polytopic LPV approach and the ultra-local model. The ultra-local model is used to update the cornering stiffness and, in this way, improve the performance level of the LPV observer. Then, the resulted observer is exploited during the LPV-based lateral control design. The improved lateral control can adapt to different circumstances such as low adhesion coefficient. The proposed observer and LPV controller is implemented and tested in MATLAB/Simulink environment and using the high-fidelity simulation software, CarMaker.

关键词： Adaptation models control design Heuristic algorithms Software algorithms Estimation Observers Mathematical models Software MATLAB Autonomous vehicles

来源：评论

学校读者我要写书评

暂无评论

Digital twin assisted workpiece referencing for compensating the stock deviation of casted parts

引用

Procedia CIRP 2023年 118卷 288-293页

作者： Tamás Cserteg Gábor Erdős Gergely Horváth András Kovács EPIC Centre of Excellence in Production Informatics and Control Institute for Computer Science and Control (SZTAKI) Budapest H-1111 Hungary Faculty of Informatics Eötvös Loránd University Budapest H-1117 Hungary Department of Manufacturing Science and Technology Budapest University of Technology and Economics Budapest H-1111 Hungary

Stock deviation of casted parts needs to be handled by adapting the CNC machining code to every new batch. Usually, human experts deal with this workpiece referencing task, but the demand for automation is expressed by the industry. This paper introduces a Digital Twin (DT) supported workpiece referencing method, implemented in the following steps: building the DT of the CNC machining cell, loading measurements of the casted part into the DT, solving the workpiece referencing problem as a convex optimization problem, and generating the compensated CNC code. The proposed approach is illustrated in a case study from the automotive industry.

关键词： Machining Workpiece referencing Digital Twin Convex optimization

来源：评论

学校读者我要写书评

暂无评论

Distribution-Free Inference for the Regression Function of Binary Classification

arXiv

引用

arXiv 2023年

作者： Tamás, Ambrus Csáji, Balázs Csanád SZTAKI The Institute for Computer Science and Control ELKH: Eötvös Loránd Research Network 13-17 Kende utca Budapest1111 Hungary Budapest Hungary

One of the key objects of binary classification is the regression function, i.e., the conditional expectation of the class labels given the inputs. With the regression function not only a Bayes optimal classifier can be defined, but it also encodes the corresponding misclassification probabilities. The paper presents a resampling framework to construct exact, distribution-free and non-asymptotically guaranteed confidence regions for the true regression function for any user-chosen confidence level. Then, specific algorithms are suggested to demonstrate the framework. It is proved that the constructed confidence regions are strongly consistent, that is, any false model is excluded in the long run with probability one. The exclusion is quantified with probably approximately correct type bounds, as well. Finally, the algorithms are validated via numerical experiments, and the methods are compared to approximate asymptotic confidence ellipsoids. Copyright © 2023, The Authors. All rights reserved.

关键词： Numerical methods

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：