The Rice Wireless Localization Toolkit

The Rice Wireless Localization (RWL) Toolkit can be used to estimate the position of a device based on the signal strength of nearby 802.11 wireless access points. While prior systems have required significant investments of human labor to build a detailed signal map, our system can be trained by spending less than one minute per office or region, and can then localize the device with very high accuracy after only two or three signal strength measurements.

The RWL was evaluated in Duncan Hall, a large office building with over 200 offices. After about 28 man-hours of training time, the RWL was able to localize a device to the precise, correct location in over 95% of our attempts, across the entire building. Even in the most pathological cases, it almost never localized the device any more distant than to a neighboring office. Furthermore, with a brief calibration period, the RWL can be adapted to work with previously unknown user hardware, and it is robust against a variety of untrained time-varying phenomena.


Related publications


If you have any questions about the RWL toolkit, please contact Andreas Haeberlen, Kostas Bekris, Prof. Dan Wallach, or Prof. Lydia Kavraki.


The following people contributed to this project (in alphabetical order): We are grateful for the generous financial support of Microsoft and Schlumberger. The RWL toolkit is dedicated to the memory of Andrew M. Ladd (1978-2007), who played an enormous role in its design, implementation, and evaluation.