Performance analysis of wireless intruder geolocation in campus wireless networks

Wireless intruder geolocation techniques using received signal strength (RSS) measurements from an array of sensors are very attractive for campus wireless network security because of their low cost implementation and simplicity. Both precision in RSS measurements and the relative geometry between RF sensors and the wireless intruder can affect the geolocation accuracy. This paper presents a comparative performance analysis of three RSS-based geolocation algorithms for non-cooperative wireless intruders in terms of their root mean square errors in wireless campus networks. A comparison study is established in a RF sensor grid and circular sensor geometries to fit general campus topology. Geolocation algorithm accuracy is examined for these particular geometric configurations of RF sensors with respect to various wireless intruder locations in and around a campus area. The results show that certain geolocation algorithms perform much better than others when wireless intruders are inside a campus wireless network, whereas the performance of the same geolocation algorithms is turned around when wireless intruders are outside a campus wireless network. As a result, a high accuracy solution to wireless intruder geolocation in campus wireless networks is discovered and proposed. MATLAB-based numerical simulations were used to evaluate the performance of these algorithms for various scenarios and parameters.