Applying Edge AI towards Deep-learning-based Monocular Visual Odometry Model for Mobile Robotics

Frederico Martins de Sousa, Mateus Silva, Ricardo Oliveira

2022

Abstract

Visual odometry is a relevant problem considering mobile robotics. While intelligent robots can provide mapping and location tasks with a multitude of sensors, it is interesting to evaluate the ability to create models using less information to create similar information. While traditional approaches consider computer vision aspects of proposing solutions, they lack the application of modern perspectives as edge computing and deep learning. This text assesses the problem of evaluating the usage of deep-learning-based visual odometry models in mobile robotics. We expect mobile robots to have embedded computers with limited computing technologies, so we approach this problem through the Edge AI perspective. Our results displayed an improvement of the model considering previous results. Also, we profile the performance of hardware candidates to perform this task in mobile edge devices.

Paper Citation

in Harvard Style

Martins de Sousa F., Silva M. and Oliveira R. (2022). Applying Edge AI towards Deep-learning-based Monocular Visual Odometry Model for Mobile Robotics. In Proceedings of the 24th International Conference on Enterprise Information Systems - Volume 2: ICEIS, ISBN 978-989-758-569-2, pages 561-568. DOI: 10.5220/0011071600003179

in Bibtex Style

@conference{iceis22,
author={Frederico Martins de Sousa and Mateus Silva and Ricardo Oliveira},
title={Applying Edge AI towards Deep-learning-based Monocular Visual Odometry Model for Mobile Robotics},
booktitle={Proceedings of the 24th International Conference on Enterprise Information Systems - Volume 2: ICEIS,},
year={2022},
pages={561-568},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011071600003179},
isbn={978-989-758-569-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 24th International Conference on Enterprise Information Systems - Volume 2: ICEIS,
TI - Applying Edge AI towards Deep-learning-based Monocular Visual Odometry Model for Mobile Robotics
SN - 978-989-758-569-2
AU - Martins de Sousa F.
AU - Silva M.
AU - Oliveira R.
PY - 2022
SP - 561
EP - 568
DO - 10.5220/0011071600003179