PENENTUAN POSISI OBJEK BERDASARKAN GLOBAL SYSTEM FOR MOBILE COMMUNICATION (GSM) MENGGUNAKAN METODE NA�VE BAYES

• Main Authors: , Hani Rubiani, , Widyawan, S.T., M. Sc., Ph.D.
• Format: Theses and Dissertations NonPeerReviewed
• Published: [Yogyakarta] : Universitas Gadjah Mada 2012
• Subjects: ETD
• Online Access: https://repository.ugm.ac.id/100567/; http://etd.ugm.ac.id/index.php?mod=penelitian_detail&sub=PenelitianDetail&act=view&typ=html&buku_id=57089
• Institution: Universitas Gadjah Mada

Most researches in indoor localization are based on the use of short-range signals, e.g. WiFi, Bluetooth, ultra sound, and infrared. This research discusses indoor localization using the Global System for Mobile Communication (GSM). The use of GSM has many advantages and works during electricity outage. The system can be used in vast area coverage and requires no information about the position of radio transmitter. The localization method uses Receive Signal Strength (RSS) GSM fingerprinting. The method has advantages which does not need information of the position of radio transmitter. The localization stage uses Naive Bayes (NB) method. Afterwards, the results will be compared with Nearest Neighbour (NN) method. Indoor localization is conducted on 3rd floor corridor of Department of Electrical Engineering and Information Technology (JTETI), UGM with total area of ±302 m2. The experiments are performed in 5 scenarios. The first scenario is the measurement of RSS fingerprint with 2 m2 grid size. The second scenario is the measurement of RSS fingerprint with 1 m2 grid size. Both scenarios employ 3 Cell-ID data from XL Axiata GSM Provider. The third scenario is similar to the first scenario with different data set. Fourth scenario employs RSS fingerprint with 1 m2 grid size and use up to 4 Cell-ID of the Telkomsel GSM Provider. The fifth scenario measures RSS fingerprint with 1 m2 grid size using 3 Cell-ID from XL Axiata GSM Provider and 3 Access Point (AP) that is attached on the 3rd floor of Department of Electrical Engineering and Information Technology (JTETI) UGM building. The results show that the fifth scenario has the highest accuracy. The scenario integrates GSM RSS fingerprint data and IEEE 802.11g RSS fingerprint. The use of Nearest Neighbour gives minimum average error of 5.39 meters. Minimum average error of 5.11 meters is obtained using Naive Bayes (NB) method.