LOG-a-TEC Testbed outdoor localization using BLE beacons

Abstract

While the Global Positioning System (GPS) provides high accuracy it places a significant strain on the device’s battery. In search of alternative techniques for outdoor localization, several approaches have been explored and recently Bluetooth Low Energy (BLE) is becoming a viable alternative to GPS for outdoor localization. Despite its popularity, access to open-source datasets for outdoor localization is limited. In this paper, we present a new openly available BLE fingerprint-based localization dataset that has been collected on LOG-a-TEC testedbed at the Jožef Stefan Institute, Ljubljana, Slovenia. The presented dataset was also used to develop a machine learning model that is capable of correctly classifying fingerprints with an average F1-score of 96.1%. We also provide insight into the importance of each node to the performance of the model. Although fingerprint-based localization is proving to be a robust alternative to GPS, we also show how changes in the environment can negatively impact the localization performance.