A Particle Filter Approach to WiFi Target Localization

This paper presents a particle filter approach to solving radio source localization using only received signal strength indicator (RSSI) measurements. It uses a model that exploits the behavior of wireless signals in free space and obtains position estimates from the signal strength. These noisy position estimates are then used in a particle filter which estimates the posterior distribution of the radio source location. Simulation results demonstrate how the method works with one or more targets and show that the error quickly stabilizes to below 200 meters after 100 readings. Experiments conducted on two sets of flight data show that the error stabilizes to below 350 meters in a duration of a few minutes. I.