Artículo de revista
Using convolutional neural networks in robots with limited computational resources: Detecting NAO robots while playing soccer
Fecha
2018Registro en:
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Volumen 11175 LNAI, 2018
16113349
03029743
10.1007/978-3-030-00308-1_2
Autor
Cruz, Nicolás
Lobos-Tsunekawa, Kenzo
Ruiz del Solar, Javier
Institución
Resumen
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%.