Feature extraction and radar track classification for detecting UAVs in civillian airspace

With the rapidly growing use of commercial Unmanned Aerial Vehicles (UAVs), integrating civilian UAVs into the controlled airspace seems inevitable. We investigate the problem of associating correct labels to different radar tracks, specifically to distinguish UAV tracks among others (such as aircraft and birds). To this end, three plausible civilian applications involving UAVs are proposed and studied. Then, for each application, a number of UAV tracks are simulated and merged into an existing dataset of real aircraft and bird tracks. We show that, with a chosen set of track features, the simulated UAV tracks are correctly labeled with 99% accuracy.