Recognition of Drone Formation Intentions Using Supervised Machine Learning

Drones are becoming a major element in defense applications, geographic surveillance, delivery of packages and their uses are expanding. Drone activity detection and identification have become an important research subject. An even more challenging problem is recognizing the intentions of a group of drones. Their intention may not be obvious, which might impose a security threat in several instances. Recognizing the targeted plan of a group of drones is the subject of study in this paper. We focus on identifying the formation a group of drones is trying to achieve. We predict the formation during the transition phase from one formation to another using softmax regression. We test several feature vector designs and present our results.