“…Due to the smaller size of most UAVs, wideband, high frequency, expensive radars are required for the accurate detection and tracking of mini UAVs [23,24,52,54], which increases the overall cost of the detection and localization system. The studies of [24,25,28,54,57,58] employed the principal component analysis (PCA) [24], convolutional neural networks (CNN) [23,28,51,54], long short-term memory (LSTM) [28], and support vector machines (SVM) [57,58] techniques for the processing of extracted features from radar signals such as micro-doppler spectrogram [23,28,54,57,58] and rangedoppler signature [24] for the classification of drones. Recently authors in [13] used the hierarchical learning approach for the detection of the presence, type, and flight trajectory of a UAV.…”