Vírus: Localização colaborativa para sistemas de posicionamento interiores

Carla Margarida da Silva Marreiros

Key information

Authors:

Carla Margarida da Silva Marreiros (Carla Margarida da Silva Marreiros)

Supervisors:

João Paulo Salgado Arriscado Costeira (João Paulo Salgado Arriscado Costeira); João Pedro Castilho Pereira Santos Gomes (João Pedro Castilho Pereira Santos Gomes)

Published in

November 23, 2020

Abstract

The work done during this dissertation has as starting point an indoor localization system developed by Thales. It uses Open Beacon based sensors and relies on RSS measurements. This system was already the subject of study of a previous dissertation. In that one, the methodologies originally used were described and others were proposed that proved to achieve better results, based on a convex optimization algorithm. In the original system (proposed by Thales) a Moving Average Filter(MAF) was applied to the measurements of RSS followed by the Friis Formula. This proved to be insufficient to determine the correct location of a target (mobile antenna). For this, in this dissertation a different method from the ones already used is studied, which also has as starting point the RSS measurements between all pairs of sensors. Given that the processing of these is done offline and not in real time, what was intended was not the exact location of the target, but the outline of the trajectory it described over time. Among the used methodologies, there are filters resembling the Kalman ones and others, MDS algorithm and Procrustes Analysis. Other processes were discarded as they proved to be unnecessary and/or unsuitable. To obtain the needed data (RSS measurements) many setups of the anchors in several different physical scenarios were examined. The selected setup to illustrate the chosen methodologies was one of the ones that produced the best results, showing the preferred features to this kind of operation.