Analysis and evaluation of Wi-Fi indoor positioning systems using smartphones

This paper attempts to analyze the main algorithms used in Machine Learning applied to the indoor location. New technologies are facing new challenges. Satellite positioning has become a typical application of mobile phones, but stops working satisfactorily in enclosed spaces. Currently there is a p...

Descripción completa

Detalles Bibliográficos
Autor: Hinojosa Muñoz, David
Tipo de recurso: tesis de maestría
Fecha de publicación:2016
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2117/96676
Acceso en línea:https://hdl.handle.net/2117/96676
Access Level:acceso abierto
Palabra clave:Wireless LANs
Machine learning
Indoor positioning
Wi-Fi
Smartphone
Xarxes locals sense fil Wi-Fi
Aprenentatge automàtic
Àrees temàtiques de la UPC::Enginyeria de la telecomunicació
Descripción
Sumario:This paper attempts to analyze the main algorithms used in Machine Learning applied to the indoor location. New technologies are facing new challenges. Satellite positioning has become a typical application of mobile phones, but stops working satisfactorily in enclosed spaces. Currently there is a problem in positioning which is unresolved. This circumstance motivates the research of new methods. After the introduction, the first chapter presents current methods of positioning and the problem of positioning indoors. This part of the work shows globally the current state of the art. It mentions a taxonomy that helps classify the different types of indoor positioning and a selection of current commercial solutions. The second chapter is more focused on the algorithms that will be analyzed. It explains how the most widely used of Machine Learning algorithms work. The aim of this section is to present mathematical algorithms theoretically. These algorithms were not designed for indoor location but can be used for countless solutions. In the third chapter, we learn gives tools work: Weka and Python. the results obtained after thousands of executions with different algorithms and parameters showing main problems of Machine Learning shown. In the fourth chapter the results are collected and the conclusions drawn are shown.