UAV Classification Utilizing Radar Digital Twins
Ahmed N. Sayed, Omar M. Ramahi, George Shaker
Abstract
The potential dangers of the unauthorized use of Unmanned Air Vehicles (UAVs) have made remote detection and classification crucial. Radar detection systems are preferred as they operate in all weather situations and during any time. Identification of UAVs threats is aided by knowledge of the number of detected UAVs and their directions. In this paper, a digital twin of a Multiple-Input Multiple-Output (MIMO) radar is used to detect a number of CAD replicas of various UAVs, and enable their simultaneous classification. Rather than resorting to complex measurement campaigns, a full-wave electromagnetic CAD tool was used to generate the digital twins, and the radar datasets which are then fed into machine learning classifiers. The proposed approach will enable antenna and radar system researchers to investigate an unprecedented plurality of possible scenarios when it comes to the use of radars for UAV detection and classification.