Conventional Wi-Fi location estimation techniques using radio fingerprinting typically require a lengthy initial site survey. It is suggested that the lengthy site survey is a barrier to adoption of the radio fingerprinting technique. This research investigated two methods for reducing or eliminating the site survey and instead build the radio map on-the-fly. The first approach utilized a deterministic algorithm to predict the user's location near each access point and subsequently construct a radio map of the entire area. This deterministic algorithm performed only fairly and only under limited conditions, rendering it unsuitable for most typical real-world deployments. Subsequently, a probabilistic algorithm was developed, derived from a robotic mapping technique called simultaneous localization and mapping. The standard robotic algorithm was augmented with a modified particle filter, modified motion and sensor models, and techniques for hardware-agnostic radio measurements (utilizing radio gradients and ranked radio maps). This algorithm performed favorably when compared to a standard implementation of the radio fingerprinting technique, but without needing an initial site survey. The algorithm was also reasonably robust even when the number of available access points were decreased.
| Identifer | oai:union.ndltd.org:GATECH/oai:smartech.gatech.edu:1853/28219 |
| Date | 07 April 2009 |
| Creators | Lim, Yu-Xi |
| Publisher | Georgia Institute of Technology |
| Source Sets | Georgia Tech Electronic Thesis and Dissertation Archive |
| Detected Language | English |
| Type | Dissertation |
Page generated in 0.0018 seconds