Sybot: an adaptive and mobile spectrum survey system for wifi networks

Information on site-specific spectrum characteristics is essential to evaluate and improve the performance of wireless networks. However, it is usually very costly to obtain accurate spectrum-condition information in heterogeneous wireless environments. This paper presents a novel spectrum-survey system, called Sybot (Spectrum survey robot), that guides network engineers to efficiently monitor the spectrum condition (e.g., RSS) of WiFi networks. Sybot effectively controls mobility and employs three disparate monitoring techniques - complete, selective, and diagnostic - that help produce and maintain an accurate spectrum-condition map for challenging indoor WiFi networks. By adaptively triggering the most suitable of the three techniques, Sybot captures spatio-temporal changes in spectrum condition. Moreover, based on the monitoring results, Sybot automatically determines several key survey parameters, such as site-specific measurement time and space granularities. Sybot has been prototyped with a commodity IEEE 802.11 router and Linux OS, and experimentally evaluated, demonstrating its ability to generate accurate spectrum-condition maps while reducing the measurement effort (space, time) by more than 56%.