Robotics Research Lab

Center for Robotics and Embedded Systems



Home
Bio & CV
Publications
Movies


Player, Stage and Gazebo
Radish

Localization using WiFi Signal Strength

As WiFi access points become increasingly prevalent in the urban environment, it is natural to ask if these access points may be exploited for purposes other than communication. In this project, we are exploring the use of WiFi signal strength information as a cue for localizing mobile robots. The basic process is as follows.

First, we construct a WiFi signal-strength map of the environment using a combination of sampling and interpolation. Samples are gathered by one or more robots; these robots are equipped with scanning laser range finders and use a standard MCL algorithm for localization [?]. The figure below, for example, shows the WiFi signal-strength map generated for the second floor of the USC SAL building (there are four access points on this floor, only three of which are shown in the signal-strength map).


Note that WiFi signal strength does not follow a parametric (inverse-square) model in indoor environments.

Second, we construct a sensor model for WiFi signal strength, taking into account sensor noise and environmental variation; we expect the signal strength map to change somewhat as people move through the environment, doors are opened or closed, and so on.

Equipped with both the signal strength map and model, robots employ a standard MCL algorithm to localize themselves. The figure below, for example, shows a particle filter converging on the correct robot pose (green dots show the WiFi-based estimate).



Click here for an animation of this experiment. Empirically, we have found that robots may be localized to within 50cm using only WiFi signal strength and odometry.

References

[howard_fsr03a] Andrew Howard, Sajid Siddiqi and Gaurav S Sukhatme, "An Experimental Study of Localization Using Wireless Ethernet", Proceedings of the 4th International Conference on Field and Service Robotics (FSR'03), pages , Lake Yamanaka, Japan, July 2003 (details)

[howard_siddiqi_icar03a] Sajid Siddiqi, Gaurav S. Sukhatme and Andrew Howard, "Experiments in Monte-Carlo Localization using WiFi Signal Strength", Proceedings of the International Conference on Advanced Robotics, pages , Coimbra, Portugal, Jul 2003 (details)

Acknowledgements

This work is sponsored in part by DARPA grants DABT63-99-1-0015 and 5-39509-A (via UPenn) under the Mobile Autonomous Robot Software (MARS) program.