Computer Science Colloquia
Monday, April 22, 2013
Advisor: Kamin Whitehouse
Attending Faculty: Jack Stankovic (Chair); Mary Lou Soffa and Gabriel Robins
2:00 PM, Rice Hall, Rm. 242
PhD Qualifying Exam Presentation
RF Doormat - Room-level tracking of residents in homes using RFID
Low power wireless localization systems can track people with accuracy up to approximately a meter in indoor spaces. These coordinate based localization techniques are however too coarse-grained to perform room-level tracking of occupants in homes, because tracking people's movement across rooms requires very precise localization near the doorways. In this thesis, we present the RF Doormat - a conceptualization, a set of guidelines and a deployment tool that enable rapid and predictable room-level tracking of residents in homes using RFID. To evaluate our system, we performed controlled experiments of 580 doorway crossings over 11 different doorways. We also analyze our system for an in-situ deployment resulting in 2550 doorway crossings in data collected for 77 hours. Results indicate that our system can perform room-level tracking with an accuracy of up to 98%.