检索结果-内蒙古大学图书馆

Improvements and Detailed Evaluation of Ground Obstacle Position, Size and Orientation Estimation ⁎

IFAC-PapersOnLine 2020年第2期53卷 14736-14741页

作者： Peter Bauer Systems and Control Laboratory Institute for Computer Science and Control (SZTAKI) Budapest Hungary

This paper improves and evaluates in detail a previous work of the author dealing with position, size and orientation estimation of fixed ground obstacles in aircraft sense and avoid. The improvement is the better conditioning of the system of equations by the shift of one known variable. Detailed evaluation means Monte-Carlo simulation and comparison of the disc and line-based parameter estimation methods considering also non-straight own aircraft trajectories. This is the focal point of the article as only straight trajectory results were compared until now. The mean estimation errors (in percentages) and their standard deviations with the two different methods are compared. Finally, the line-based method was better applicable for rectangular objects as expected.

关键词： Obstacle detection Ground obstacle Position estimation Size estimation

来源：评论

学校读者我要写书评

暂无评论

Voting-based Fault Detection for Aircraft Position Measurements with Dissimilar Observations ⁎

引用

IFAC-PapersOnLine 2020年第2期53卷 14724-14729页

作者： Tamas Grof Peter Bauer Systems and Control Laboratory Institute for Computer Science and Control (SZTAKI) Budapest Hungary

In this article a fault detection algorithm for aircraft position measurements i proposed using redundant sensor information during landing scenarios. This work was developed in the framework of the VISION EU H2020 research project. The aircraft’s position can be determined via instrumental landing system, GPS and camera measurements. Considering these three sources a two out of three voting logic can be developed. After transforming the measured data sets to a common format two different methods are constructed to execute voting. The first is simple and well known thresholding where the measured position values are compared pairwise and thresholding violations registered. As dissimilar data noise strengths can make thresholdin unreliable the second method proposed by the authors is supplemented with an additional statistical evaluation where the measurements undergo a two-sample Z-test. Both methods were evaluated off-line with Monte-Carlo computer simulation. The tests showed that the proposed statistical method outperforms the straightforward thresholding approach.

关键词： Fault detection Fault isolation Sensor failures Threshold logic

来源：评论

学校读者我要写书评

暂无评论

Real-time Vehicle Localization and Pose Tracking in High-Resolution 3D Maps

Real-time Vehicle Localization and Pose Tracking in High-Res...

引用

European Signal Processing Conference (EUSIPCO)

作者： Örkény Zováthi Balázs Pálffy Csaba Benedek Machine Perception Research Laboratory Institute for Computer Science and Control (SZTAKI) Budapest Hungary Faculty of Information Technology and Bionics Pázmány Péter Catholic University Budapest Hungary

ISBN: (数字)9789082797091

ISBN: (纸本)9781665467995

In this paper we introduce a novel approach for accurate self-localization and pose tracking for Lidar and GPS-equipped autonomous vehicles (AVs) in high-density (more than 5000 points/m 2 ) 3D localization maps obtained through Mobile Laser Scanning (MLS). Our solution consist of two main steps: First, starting from a poor GPS-based initial position, we estimate the 3DoF pose (planar position and yaw orientation) of the ego vehicle by aligning its sparse (50-500 points/m 2 ) Lidar point cloud measurements to the MLS prior map, using a novel approach of matching static landmark objects of the scene. Second, to effectively deal with the lack of pairable objects in certain time frames (e.g. due to scene segments occluded by a large moving tram), we track the estimated 3DoF pose of the AVs by a Kalman filter. Comperative test are provided on roads with heavy traffic in downtown city areas with large (5-10 meters) GPS positioning errors. The proposed approach is able to reduce the location error of the vehicle by one order of magnitude and keep the yaw angle error around 1° during its whole trajectory without considerable drift, while running in real-time (20-25 Hz).

关键词： Location awareness Maximum likelihood estimation Laser radar Three-dimensional displays Real-time systems Kalman filters Vehicle dynamics

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Learning Based Resource Management for CAR T-Cell Therapies

引用

Procedia CIRP 2024年 125卷 154-159页

作者： Szabolcs Szentpéteri Krisztián B. Kis Péter Egri Carmen Sanges Sophia Danhof Katrin Mestermann Michael Hudecek Sergio Navarro Velázquez Manel Juan Balázs Cs. Csáji Institute for Computer Science and Control (SZTAKI) Hungarian Research Network (HUN-REN) Kende str 13-17 Budapest 1111 Hungary Universitätsklinikum Würzburg Versbacher Straße 5 Würzburg 97078 Germany Hospital Clínic Barcelona Villarroel 170 - Escala 3 Planta 1 Barcelona 08036 Spain Institute of Mathematics Eötvös Loránd University (ELTE) Budapest 1117 Hungary

This paper focuses on optimizing resource management strategies in chimeric antigen receptor (CAR) T-cell therapies using reinforcement learning (RL). CAR T-cell therapy is an innovative and promising treatment within the field of immunotherapy, which is based on the isolation of patients’ own T-cells, genetically modifying these cells to express a CAR for tumor recognition, cultivating and expanding these T cells and infusing them back to the patient. These therapies require several reusable but scarce resources, including special equipment, such as the bioreactor that is used to expand CAR T cells, and medical resources, such as the hospital staff, e.g., doctors and nurses. Considering the stochastic nature of medical procedures and the production of CAR T cells, the scheduling of the therapies and the efficient allocation of the required resources pose a significant challenge inside the hospital environment. Here, we propose a derivative-free policy gradient algorithm that utilizes a simulation model of the therapy, built using real-world data, to obtain efficient control policies for the resource management problem. The proposed method is designed to minimize the expected number of deviations from the therapy protocol as well as the expected overall completion time of the therapies for every patient. The effectiveness of the proposed resource management approach is demonstrated via simulation experiments.

关键词： patient scheduling reinforcement learning therapy simulation CAR T-cell therapy

来源：评论

学校读者我要写书评

暂无评论

Decentralized and prioritized algorithm for AGV fleet management

引用

IFAC-PapersOnLine 2021年第1期54卷 98-103页

作者： Tímea Tamási Tamás Kis ELTE Eötvös Loránd University Doctoral School of Mathematics Budapest Hungary ELKH SZTAKI Institute for Computer Science and Control 1111 Budapest Kende u. 13-17. Hungary ELKH SZTAKI Institute for Computer Science and Control 1111 Budapest Kende u. 13-17. Hungary

In this paper we propose an algorithm based on prioritization for coordinating autonomously guided vehicles (AGVs) in a grid/mesh topology, where the tasks arrive online into the system. The method combines a decentralized control mechanism based on prioritization with an optimization subroutine to resolve eventually occurring deadlock situations. The merits of the method are evaluated on a set of benchmark instances.

关键词： autonomously guided vehicles decentralized control cooperative navigation prioritized algorithm optimization

来源：评论

学校读者我要写书评

暂无评论

Detection and Tracking of MAVs Using a Rosette Scanning Pattern LiDAR

arXiv

引用

arXiv 2024年

作者： Gazdag, Sándor Möller, Tom Keszler, Anita Majdik, András L. HUN-REN SZTAKI Institute for Computer Science and Control Kende u. 13-17 Budapest1111 Hungary Budapest University of Technology and Economics Kálmán Kandó Doctoral School of Transportation and Vehicle Engineering Műegyetem rkp. 3. Budapest1111 Hungary Hungary

The use of commercial Micro Aerial Vehicles (MAVs) has surged in the past decade, offering societal benefits but also raising risks such as airspace violations and privacy concerns. Due to the increased security risks, the development of autonomous drone detection and tracking systems has become a priority. In this study, we tackle this challenge, by using non-repetitive rosette scanning pattern LiDARs, particularly focusing on increasing the detection distance by leveraging the characteristics of the sensor. The presented method utilizes a particle filter with a velocity component for the detection and tracking of the drone, which offers added re-detection capability. A Pan-Tilt platform is utilized to take advantage of the specific characteristics of the rosette scanning pattern LiDAR by keeping the tracked object in the center where the measurement is most dense. The detection capabilities and accuracy of the system are validated through indoor experiments, while the maximum detection distance is shown in our outdoor experiments. Our approach achieved accuracy on par with the state-of-the-art indoor method while increasing the maximum detection range by approximately 80% beyond the state-of-the-art outdoor method. © 2024, CC BY.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

Object localization utilizing 3D point cloud clustering approach 53

Object localization utilizing 3D point cloud clustering appr...

引用

53rd CIRP Conference on Manufacturing Systems, CMS 2020

作者： Horváth, Gergely Erdos, Gábor SZTAKI: Institute for Computer Science and Control Budapest1111 Hungary Department of Manufacturing Science and Engineering Budapest University of Technology and Economics Budapest1111 Hungary

In this paper we present a novel method to recognize multiple workpieces on a flat surface, based on a 3D point cloud and a prebuilt pose database. First, the method identifies the main surface, which is the base of every following steps. The surface determines the plane stretched by X and Y axis, establishing the orientation of the objects on it and helping their isolated recognition. Isolation of objects is followed by determining their specific type and exact pose, based on the prebuilt database. The presented method is demonstrated by recognizing workpieces and tool in an assembly cell. © 2020 The Authors.

关键词： Object recognition

来源：评论

学校读者我要写书评

暂无评论

Towards Understanding Worldwide Cross-Cultural Differences in Implicit Driving Cues: Review, Comparative Analysis, and Research Roadmap

Towards Understanding Worldwide Cross-Cultural Differences i...

引用

International Conference on Intelligent Transportation

作者： Yongqi Dong Chang Liu Yiyun Wang Zhe Fu Delft University of Technology Delft The Netherlands RWTH Aachen University Aachen Germany Department of Networked Systems and Services Budapest University of Technology and Economics Budapest Hungary Machine Perception Research Laboratory of HUN-REN Institute for Computer Science and Control (SZTAKI) Budapest Hungary University of California Berkeley CA United States

ISBN: (数字)9798331505929

ISBN: (纸本)9798331505936

Recognizing and understanding implicit driving cues across diverse cultures is imperative for fostering safe and efficient global transportation systems, particularly when training new immigrants holding driving licenses from culturally disparate countries. Additionally, it is essential to consider cross-cultural differences in the development of Automated Driving features tailored to different countries. Previous piloting studies have compared and analyzed cross-cultural differences in selected implicit driving cues, but they typically examine only limited countries. However, a comprehensive worldwide comparison and analysis are lacking. This study conducts a thorough review of existing literature, online blogs, and expert insights from diverse countries to investigate cross-cultural disparities in driving behaviors, specifically focusing on implicit cues such as non-verbal communication (e.g., hand gestures, signal lighting, honking), norms, and social expectations. Through comparative analysis, variations in driving cues are illuminated across different cultural contexts. Based on the findings and identified gaps, a research roadmap is proposed for future research to further explore and address these differences, aiming to enhance intercultural communication, improve road safety, and increase transportation efficiency on a global scale. This paper presents the pioneering work towards a comprehensive understanding of the implicit driving cues across cultures. Moreover, this understanding will inform the development of automated driving systems tailored to different countries considering cross-cultural differences.

关键词： Hands Training Reviews Velocity control Transportation Lighting Road safety Cultural differences Global communication Driver behavior

来源：评论

学校读者我要写书评

暂无评论

Evaluating transferability of BERT models on uralic languages

arXiv

引用

arXiv 2021年

作者： Ács, Judit Lévai, Dániel Kornai, András SZTAKI Institute for Computer Science and Control Department of Digital Humanities Eötvös Loránd University Hungary

Transformer-based language models such as BERT have outperformed previous models on a large number of English benchmarks, but their evaluation is often limited to English or a small number of well-resourced languages. In this work, we evaluate monolingual, multilingual, and randomly initialized language models from the BERT family on a variety of Uralic languages including Estonian, Finnish, Hungarian, Erzya, Moksha, Karelian, Livvi, Komi Permyak, Komi Zyrian, Northern Sámi, and Skolt Sámi. When monolingual models are available (currently only et, fi, hu), these perform better on their native language, but in general they transfer worse than multilingual models or models of genetically unrelated languages that share the same character set. Remarkably, straightforward transfer of high-resource models, even without special efforts toward hyperparameter optimization, yields what appear to be state of the art POS and NER tools for the minority Uralic languages where there is sufficient data for finetuning. © 2021, CC0.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Observer design with performance guarantees for vehicle control purposes via the integration of learning-based and LPV approaches

Observer design with performance guarantees for vehicle cont...

引用

IEEE Intelligent Vehicles Symposium Workshops (IV Workshops)

作者： Dániel Fényes Tamás Hegedűs Balázs Németh Péter Gáspár Institute for Computer Science and Control (SZTAKI) Eötvös Loránd Research Network (ELKH) Budapest Hungary

ISBN: (纸本)9781665479226

The paper proposes an enhanced observer design method for autonomous vehicles, with which the unmeasurable states in vehicle and chassis motion can be estimated. The novelty of the method is that the learning-based observer and the linear parameter varying (LPV) observer in a joint observer design structure are incorporated, which results in an improved performance level on the estimation error. Nevertheless, the proposed design method is able to guarantee the limitation of the estimation error, even if the error of the learning-based observer under all scenarios cannot be verified. Thus, the proposed method handles the main disadvantage of the learning-based approaches, i.e. the lack of performance guarantees, while their advantages, i.e. the improved observation performance in the operation of the observer are taken. The proposed method is applied on a lateral path following control problem, where the goal of the observer is to provide an accurate lateral velocity signal for the vehicle. The effectiveness of the method is illustrated through simulation examples on high- fidelity vehicle dynamic simulator CarSim.

关键词： Estimation error Uncertain systems Systematics Design methodology Conferences Roads Observers

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：