Using convolutional neural networks in robots with limited computational resources: Detecting NAO robots while playing soccer
Author
dc.contributor.author
Cruz, Nicolás
Author
dc.contributor.author
Lobos-Tsunekawa, Kenzo
Author
dc.contributor.author
Ruiz del Solar, Javier
Admission date
dc.date.accessioned
2019-05-31T15:21:10Z
Available date
dc.date.available
2019-05-31T15:21:10Z
Publication date
dc.date.issued
2018
Cita de ítem
dc.identifier.citation
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Volumen 11175 LNAI, 2018
Identifier
dc.identifier.issn
16113349
Identifier
dc.identifier.issn
03029743
Identifier
dc.identifier.other
10.1007/978-3-030-00308-1_2
Identifier
dc.identifier.uri
https://repositorio.uchile.cl/handle/2250/169520
Abstract
dc.description.abstract
The main goal of this paper is to analyze the general problem of using Convolutional Neural Networks (CNNs) in robots with limited computational capabilities, and to propose general design guidelines for their use. In addition, two different CNN based NAO robot detectors that are able to run in real-time while playing soccer are proposed. One of the detectors is based on the XNOR-Net and the other on the SqueezeNet. Each detector is able to process a robot object-proposal in ~1 ms, with an average number of 1.5 proposals per frame obtained by the upper camera of the NAO. The obtained detection rate is ~97%.