A Self-learning Fingerprinting Matching Algorithm for Indoor Wi-Fi Positioning

Guenther Retscher, Anton Stangl

Peer Reviewed

Abstract: Wi-Fi location fingerprinting is nowadays a popular indoor localization technique. The self-learning fingerprinting matching approach presented in this paper is a new strategy to reduce the labor and time consuming system training. Continuous training is performed instead of a dedicated training phase before positioning of a user is possible. Verified RSSI (Received Signal Strength Indicator) scans made by the users’ mobile device are then included into the fingerprinting database and continuously update the system. Due to this continuous system training spatial and temporal RSSI variations can be accommodated and modelled. With this approach a significant increase in localization accuracy and performance is achieved. Tests in a multi-storey office environment and a two storey residential apartment as well as a medical praxis showed matching rates of over 90 % for the localization of the mobile user in a certain room.
Published in: 2018 IEEE/ION Position, Location and Navigation Symposium (PLANS)
April 23 - 26, 2018
Hyatt Regency Hotel
Monterey, CA
Pages: 1009 - 1019
Cite this article: Retscher, Guenther, Stangl, Anton, "A Self-learning Fingerprinting Matching Algorithm for Indoor Wi-Fi Positioning," 2018 IEEE/ION Position, Location and Navigation Symposium (PLANS), Monterey, CA, April 2018, pp. 1009-1019.
Full Paper: ION Members/Non-Members: 1 Download Credit
Sign In