Abstract (EN):
This paper describes state-of-art radio frequency (RF) localization in indoors quasi-structured environments. Trilateration is a technique adopted by RF localization method which makes use of available communications subsystem. Therefore, localization hardware is supported by that sub-system and only additional software is shown to be necessary in building localization framework. This paper shows that RF localization can be applied to indoors localization of automated guided vehicles (AGVs) with improvement of localization accuracy using artificial neural networks (ANNs). ANN can be a type of transfer function emulation scheme that does not require knowledge of corresponding function properties and, with a feed-forward structure trained through error back-propagation algorithms, it can be useful to estimate AGV location through RF link signal parameter measurements.
Idioma:
Inglês
Tipo (Avaliação Docente):
Científica
Contacto:
azenha@fe.up.pt