Indoor Positioning Using WLAN Fingerprinting with Post-Processing Scheme

Abstract

Information about a person’s position is a valuable piece of context information on which many application and location services are based upon. In outdoor environments the Global Positioning System (GPS) and Assisted GPS (A-GPS) are widely used and they perform reasonably well, but they underperform when there is no clear access to the sky, i.e. in indoor environments. Most of the research conducted and solutions developed aim for real-time indoor positioning or personal tracking, but to the author’s knowledge there are not many studies on the subject of post- processing. Post-processing has many benefits over real-time solutions, like preserving battery life of a mobile device, leveraging bigger processing power, using more complex algorithms that cannot run on mobile devices, and ultimately getting better accuracy on a person’s movements tracks. In this thesis, an Indoor Positioning System (IPS) using WLAN fingerprinting with post- processing scheme is proposed. The system uses a large set of fingerprinted Received Signal Strength (RSS) collections obtained in the offline phase and references them in post-processing against data collected in the online phase. A series of field experiments have been conducted in University of Tartu’s Faculty of Mathematics and Computer Science building. The results show that with a post-processing scheme more computationally extensive algorithms can be used and better accuracy achieved than in real-time.