Deep Learning-Based Localization Approach for Autonomous Robots in the RobotAtFactory 4.0 Competition

Luan Carlos Klein; João Mendes; João Braun; Felipe Martins; André Schneider de Oliveira; Paulo Gomes Costa; Heinrich Wörtche; José Lima

doi:10.1007/978-3-031-53036-4_13

Deep Learning-Based Localization Approach for Autonomous Robots in the RobotAtFactory 4.0 Competition

Luan Carlos Klein, João Mendes, João Braun, Felipe Martins, André Schneider de Oliveira, Paulo Gomes Costa, Heinrich Wörtche, José Lima

Sensors and Smart Systems

Research output: Contribution to conference › Paper › Academic

Abstract

Accurate localization in autonomous robots enables effective decision-making within their operating environment. Various methods have been developed to address this challenge, encompassing traditional techniques, fiducial marker utilization, and machine learning approaches. This work proposes a deep-learning solution employing Convolutional Neural Networks (CNN) to tackle the localization problem, specifically in the context of the RobotAtFactory 4.0 competition. The proposed approach leverages transfer learning from the pre-trained VGG16 model to capitalize on its existing knowledge. To validate the effectiveness of the approach, a simulated scenario was employed. The experimental results demonstrated an error within the millimeter scale and rapid response times in milliseconds. Notably, the presented approach offers several advantages, including a consistent model size regardless of the number of training images utilized and the elimination of the need to know the absolute positions of the fiducial markers.

Original language	English
Pages	181-194
Number of pages	14
DOIs	https://doi.org/10.1007/978-3-031-53036-4_13
Publication status	Published - 3 Feb 2024
Event	Optimization, Learning Algorithms and Applications - OL2A 2023 - Ponta Delgada, Portugal Duration: 27 Sept 2023 → 29 Sept 2023 https://ol2a.ipb.pt/ol2a_2023.html

Conference

Conference	Optimization, Learning Algorithms and Applications - OL2A 2023
Abbreviated title	OL2A 2023
Country/Territory	Portugal
City	Ponta Delgada
Period	27/09/23 → 29/09/23
Internet address	https://ol2a.ipb.pt/ol2a_2023.html

Keywords

CNN
indoor localization
robotics competitions

Research Focus Areas Hanze University of Applied Sciences * (mandatory by Hanze)

Healthy Ageing

Research Focus Areas Research Centre or Centre of Expertise * (mandatory by Hanze)

Artificial Intelligence
Cyberfysical systems

Publinova themes

ICT and Media
Health
Technology

Access to Document

10.1007/978-3-031-53036-4_13Licence: CC BY

Best Paper Award - OL2A 2023
Klein, L. C. (Recipient), Mendes, J. (Recipient), Braun, J. (Recipient), Martins, F. (Recipient), Schneider de Oliveira, A. (Recipient), Costa, P. G. (Recipient), Wörtche, H. (Recipient) & Lima, J. (Recipient), 29 Sept 2023
Prize: Prize (including medals and awards)

File

Cite this

Klein, L. C., Mendes, J., Braun, J., Martins, F., Schneider de Oliveira, A., Costa, P. G., Wörtche, H., & Lima, J. (2024). Deep Learning-Based Localization Approach for Autonomous Robots in the RobotAtFactory 4.0 Competition. 181-194. Paper presented at Optimization, Learning Algorithms and Applications - OL2A 2023, Ponta Delgada, Portugal. https://doi.org/10.1007/978-3-031-53036-4_13

@conference{459bfee9bdb84d2187f3b239c9d94d32,

title = "Deep Learning-Based Localization Approach for Autonomous Robots in the RobotAtFactory 4.0 Competition",

abstract = "Accurate localization in autonomous robots enables effective decision-making within their operating environment. Various methods have been developed to address this challenge, encompassing traditional techniques, fiducial marker utilization, and machine learning approaches. This work proposes a deep-learning solution employing Convolutional Neural Networks (CNN) to tackle the localization problem, specifically in the context of the RobotAtFactory 4.0 competition. The proposed approach leverages transfer learning from the pre-trained VGG16 model to capitalize on its existing knowledge. To validate the effectiveness of the approach, a simulated scenario was employed. The experimental results demonstrated an error within the millimeter scale and rapid response times in milliseconds. Notably, the presented approach offers several advantages, including a consistent model size regardless of the number of training images utilized and the elimination of the need to know the absolute positions of the fiducial markers.",

keywords = "CNN, indoor localization, robotics competitions, indoor lokalisatie, robotica wedstrijden",

author = "Klein, {Luan Carlos} and Jo{\~a}o Mendes and Jo{\~a}o Braun and Felipe Martins and {Schneider de Oliveira}, Andr{\'e} and Costa, {Paulo Gomes} and Heinrich W{\"o}rtche and Jos{\'e} Lima",

note = "DBLP License: DBLP's bibliographic metadata records provided through http://dblp.org/ are distributed under a Creative Commons CC0 1.0 Universal Public Domain Dedication. Although the bibliographic metadata records are provided consistent with CC0 1.0 Dedication, the content described by the metadata records is not. Content may be subject to copyright, rights of privacy, rights of publicity and other restrictions.; Optimization, Learning Algorithms and Applications - OL2A 2023, OL2A 2023 ; Conference date: 27-09-2023 Through 29-09-2023",

year = "2024",

month = feb,

day = "3",

doi = "10.1007/978-3-031-53036-4_13",

language = "English",

pages = "181--194",

url = "https://ol2a.ipb.pt/ol2a_2023.html",

}

Klein, LC, Mendes, J, Braun, J, Martins, F, Schneider de Oliveira, A, Costa, PG, Wörtche, H & Lima, J 2024, 'Deep Learning-Based Localization Approach for Autonomous Robots in the RobotAtFactory 4.0 Competition', Paper presented at Optimization, Learning Algorithms and Applications - OL2A 2023, Ponta Delgada, Portugal, 27/09/23 - 29/09/23 pp. 181-194. https://doi.org/10.1007/978-3-031-53036-4_13

TY - CONF

T1 - Deep Learning-Based Localization Approach for Autonomous Robots in the RobotAtFactory 4.0 Competition

AU - Klein, Luan Carlos

AU - Mendes, João

AU - Braun, João

AU - Martins, Felipe

AU - Schneider de Oliveira, André

AU - Costa, Paulo Gomes

AU - Wörtche, Heinrich

AU - Lima, José

N1 - DBLP License: DBLP's bibliographic metadata records provided through http://dblp.org/ are distributed under a Creative Commons CC0 1.0 Universal Public Domain Dedication. Although the bibliographic metadata records are provided consistent with CC0 1.0 Dedication, the content described by the metadata records is not. Content may be subject to copyright, rights of privacy, rights of publicity and other restrictions.

PY - 2024/2/3

Y1 - 2024/2/3

N2 - Accurate localization in autonomous robots enables effective decision-making within their operating environment. Various methods have been developed to address this challenge, encompassing traditional techniques, fiducial marker utilization, and machine learning approaches. This work proposes a deep-learning solution employing Convolutional Neural Networks (CNN) to tackle the localization problem, specifically in the context of the RobotAtFactory 4.0 competition. The proposed approach leverages transfer learning from the pre-trained VGG16 model to capitalize on its existing knowledge. To validate the effectiveness of the approach, a simulated scenario was employed. The experimental results demonstrated an error within the millimeter scale and rapid response times in milliseconds. Notably, the presented approach offers several advantages, including a consistent model size regardless of the number of training images utilized and the elimination of the need to know the absolute positions of the fiducial markers.

AB - Accurate localization in autonomous robots enables effective decision-making within their operating environment. Various methods have been developed to address this challenge, encompassing traditional techniques, fiducial marker utilization, and machine learning approaches. This work proposes a deep-learning solution employing Convolutional Neural Networks (CNN) to tackle the localization problem, specifically in the context of the RobotAtFactory 4.0 competition. The proposed approach leverages transfer learning from the pre-trained VGG16 model to capitalize on its existing knowledge. To validate the effectiveness of the approach, a simulated scenario was employed. The experimental results demonstrated an error within the millimeter scale and rapid response times in milliseconds. Notably, the presented approach offers several advantages, including a consistent model size regardless of the number of training images utilized and the elimination of the need to know the absolute positions of the fiducial markers.

KW - CNN

KW - indoor localization

KW - robotics competitions

KW - indoor lokalisatie

KW - robotica wedstrijden

U2 - 10.1007/978-3-031-53036-4_13

DO - 10.1007/978-3-031-53036-4_13

M3 - Paper

SP - 181

EP - 194

T2 - Optimization, Learning Algorithms and Applications - OL2A 2023

Y2 - 27 September 2023 through 29 September 2023

ER -

Deep Learning-Based Localization Approach for Autonomous Robots in the RobotAtFactory 4.0 Competition

Abstract

Conference

Keywords

Research Focus Areas Hanze University of Applied Sciences * (mandatory by Hanze)

Research Focus Areas Research Centre or Centre of Expertise * (mandatory by Hanze)

Publinova themes

Access to Document

Fingerprint

Prizes

Best Paper Award - OL2A 2023

Cite this