Crowdsourced Wi-Fi Access Point Localization using Vertical Movement Detection

Precise indoor positioning requires building databases depending on applied location estimation techniques in general. This paper studies an automated framework to determine the locations of Wi-Fi access points (APs) using data from multiple mobile users. As mobile users only move vertically in certain points (e.g., stairs), the proposed framework detects vertical movements from air pressure measurements, extracts user trajectories on a target floor, and places the extracted trajectories between a pair of vertically movable points. In order to detect vertical movement, we assume the indoor trajectory starts on one floor and moves to another floor, resulting in more than 2 vertical movements. Finally, the AP locations are estimated using Wi-Fi signal strength measured along the placed user trajectories. The effectiveness of the proposed framework is verified under a practical indoor environment using an Android application.