Robust indoor positioning system in poor infrastructure environments

Abstract

In order to estimate an accurate position of user in an indoor environment, fingerprint location estimation technology has been actively studying. Fingerprint method is generally affected by the installed infrastructure environment. In this study, we propose a technique for estimating the accurate position even in a poor infrastructure environment. The proposed technique saves the received RSSI value in the user's buffer and utilizes the accumulated RSSI value when estimating the user position. Using the pedestrian dead reckoning technique, the user's trajectory is estimated, and the user's RSSI spatial pattern is created using the RSSI value received on the corresponding trajectory. And by comparing the user spatial pattern with the fingerprint DB, the position having the highest correlation is found. Using the proposed method, an experiment was conducted using only three BLE beacons in a space of about 20m X 50m. As a result of the experiment, a location estimation result of less than 3m was obtained.