MERIT: MEsh of RF sensors for Indoor Tracking

Abstract

A traditional approach to indoor tracking utilizes non-RF ranging techniques, such as infrared or ultrasound. The problem with these non-RF ranging techniques is that they do not work well when the tracking devices are buried in users' wallets or bags. As a result, there has been considerable interest in using only RF techniques for indoor tracking. Existing RF-only techniques, however, typically require a costly site survey and a floor-plan. In this paper, we present the MERIT system that we designed, implemented, and evaluated. MERIT is significantly different from existing systems in that it is pure RF-based yet it does not require a site survey nor a floor-plan. MERIT tracks users to a room granularity, and it can disambiguate neighboring rooms. This disambiguation is challenging because RF signals can traverse through walls. Also, because of indoor multipath interference, it is difficult to correlate signal strength with distance. In this work, we proposed two techniques for accurate disambiguation: spatial diversity and RF reflector. In our evaluation MERIT achieved an accuracy of 98.9%. MERIT was first conceived for a telecommunication application - intelligent telephone call routing, but it can also be used for other location-aware services