A Drone Early Warning System for Predicting Threatening Trajectories

Tonmoay Deb, Sven de Laaf, Valerio La Gatta, Odette Lemmens, Roy Lindelauf, Max van Meerten, Herwin Meerveld, Afke Neeleman, Marco Postiglione, Venkatramanan Siva Subrahmanian

Abstract: Over the last few years, there has been increasing use of drones by terror groups and in armed conflict. Several technologies have been developed to detect drone flights. However, much less work has been done on the drone threat prediction problem (DTPP): Predicting which drone trajectories are threatening and which ones are not. We propose the drone early warning system (DEWS), a framework to solve this problem. Solving DTPP early is key. Once a drone starts on its trajectory, we show that DEWS can make accurate predictions within 20–30 s of the flight with an F1-score of over 80% on data about a major European city. We study the tradeoff between earliness of predictions and accuracy. We identify the key features that ensure good predictions.

External IDs:dblp:journals/expert/DebLGLLMMNPS25