BaroDetector: Building Level Location Track Using Smartwatch

초록

Thanks to the many developments in mobile computing technology, including smartphones, we enjoy many conveniences in our daily lives. In particular, location-based services allow us to receive appropriate advertisements or receive alerts about dangerous situations. Most of these location-based services rely on GPS sensors. However, the GPS sensor consumes a lot of battery power of the smartphone. Moreover, when the user enters indoors, it is difficult to track the exact location of the user. In this paper, we propose a building-level location track using the pressure sensor on a smartwatch and the number of visible satellites from GPS data on smartphone. We detect door passing event with 97.7% accuracy by training machine learning model with pressure sensor data. The proposed method, when it determines that the user has passed through the door, it checks the trend of the number of visible satellites to determine whether user enters indoor. After that, the last GPS coordinates are fixed on the map to display the building where the user entered, and deactivate the GPS sensor to reduce battery consumption. The GPS sensor is reactivated when the user goes outdoor. The proposed method detects the user’s indoor/outdoor state with an average accuracy of 92.2%. © 2022, Korean Institute of Communications and Information Sciences. All rights reserved.

키워드