Thesis Details
Využití strojového učení pro zvýšení robustnosti určení pozice v bezdrátovém pozičním systému
This thesis describes the Sewio platform and the communication techniques of the ultra-wideband technology standard, which the platform uses to determine the position of objects. The technology is based on measuring signal arrival time intervals and multilateration using time differences. The platform generates and stores historical data from past positioning of objects. The dataset consists of sequences of position data which, in addition to the monitored environment, contain relevant signal parameters of wireless communication. A system of machine learning techniques based on Gaussian models and linear regression was implemented to classify and predict real-time position data with the goal of improving position estimation stability and robustness. The system functions as a downstream component, which accepts RTLS position data and outputs improved position estimates. The evaluation results show that the implemented system can successfully improve position stability and robustness.
Machine learning, real-time location system (RTLS), ultra-wideband (UWB), indoor positioning, classification, prediction
@mastersthesis{FITMT24508,
author = "Adam Matu\v{s}",
type = "Master's thesis",
title = "Vyu\v{z}it\'{i} strojov\'{e}ho u\v{c}en\'{i} pro zv\'{y}\v{s}en\'{i} robustnosti ur\v{c}en\'{i} pozice v bezdr\'{a}tov\'{e}m pozi\v{c}n\'{i}m syst\'{e}mu",
school = "Brno University of Technology, Faculty of Information Technology",
year = 2022,
location = "Brno, CZ",
language = "czech",
url = "https://www.fit.vut.cz/study/thesis/24508/"
}