Machine Learning Based Classification of Radar Signatures of Drones

With the current popularity of drones and UAVs, there is an urgent need to be able to classify the aerial objects with sufficient accuracy. Hence, several methods have been proposed for classification of drones and UAVs. Such methods are often based on visual sources and thus classification becomes dependent on extraneous parameters. In contrast, the use of Radar Cross Section (RCS) for drone classification shows less dependency of extraneous parameters. Radar Cross Section (RCS) is a significant radar signature that is popularly used for identification of targets. In this paper, the primary objective is to demonstrate a viable solution to classify drones based on their RCS values. During the process, various classification algorithms such as Support Vector Machines, Decision Tree, Naive Bayes Classifier, Neural Networks have been investigated for performance and accuracy.