Initial results of Radar-based classification of commercial drone carrying small payloads

This student paper shows preliminary results on the possibility of using radar systems to identify small payloads carried by a commercial drone. This can be of great interest for traffic monitoring and situational awareness in near-future scenarios when drones are expected to be used for parcel delivery and many other applications. In this work, we consider a distributed and rather small payload, approximately 92g, carried by a commercial drone flying back and forth. Different combinations of features extracted from the micro-Doppler signatures and dwell time have been analyzed, showing promising results with median accuracy around 80% and maximum accuracy approaching 90% for the optimal case.