Title :
Practical considerations for RSS RF fingerprinting based indoor localization systems
Author :
Bacak, Ahmet ; Celebi, Haluk
Author_Institution :
Bilgisayar Muhendisligi Bolumu, Gebze Yuksek Teknoloji Enstitusu, Gebze, Turkey
Abstract :
There are different location estimation approaches in order to determine the position of a mobile users precisely in indoor environments. In this study, in order to determine the position of the mobile user at the shopping malls received signal strength (RSS) RF fingerprint based approaches is considered. Wi-Fi and GSM RSS data has been collected by using LG Nexus 4 cell phone at Gebze Center shopping mall in Gebze, Kocaeli. By using the collected data, the effects of different machine learning algorithms, number of training data, number of measurement grid, and signal type on the performance of the localization systems are studied. According to the results, it is observed that, combining Wi-Fi and GSM RSS data measurements decreases the location estimation error.
Keywords :
cellular radio; indoor radio; learning (artificial intelligence); mobile radio; wireless LAN; GSM; Gebze Center shopping mall; Kocaeli; LG Nexus 4 cell phone; RF fingerprinting; RSS; Wi-Fi; indoor localization systems; location estimation; machine learning algorithms; mobile users; received signal strength; shopping malls; Conferences; GSM; Global Positioning System; IEEE 802.11 Standards; Radio frequency; Signal processing; Support vector machines; indoor localization; rf fingerprinting; the estimation of mobile user;
Conference_Titel :
Signal Processing and Communications Applications Conference (SIU), 2014 22nd
Conference_Location :
Trabzon
DOI :
10.1109/SIU.2014.6830274
