Machine learning techniques for geolocalization of Wi-Fi devices using flying Unmanned Aerial Vehicles

The purpose of this work is to develop a localization algorithm that can be used by flying UAVs to locate Wi-Fi devices on the ground. Assuming uncooperative target, we measure the RSSI on the packets that are periodically and spontaneously broadcasted by the Wi-Fi cards, and we use this information to estimate their position. GPs allowed us to reinforce the system against the noise that affects the measurements, and bayesian optimization proved an efficient way for sampling