Abstract

This research discuss about object position estimation (localization) inside the building using wireless network or IEEE 802.11g with machine learning approach. Measurement method of RSS using RSS-based fingerprint. Machine learning algorithm used to estimate the location of measuring RSS-based with Naive Bayes. The localozation conducted in third floor at Electronic Engineering and Information Technology Departement (JTETI) with 1969,68 m2 area and 5 point of access point placement. To build fingerprint map, we use 1m x 1m dimension thus formed 1983 grids. Using Net Surveyor software, received signal strength from wireless to receiver device gathered 86,980 record. Mean distance estimation error value for the localization for whole room at third floor using Naïve Bayes algorthim in offline phase of learning stage is 6.29 meter. From online phase of learning stage resulting mean distance error 7.82 meter.

Authors and Affiliations