Drones Classification by the Use of a Multifunctional Radar and Micro-Doppler Analysis

The classification of targets by the use radars has received great interest in recent years, particular defence and military applications, which development sensor systems that are able to identify classify threatening is a mandatory requirement. In specific case drones, several techniques have already been proposed and, up now, most effective technique was considered be micro-Doppler analysis used conjunction with machine learning tools. signatures usually represented form spectrogram, time–frequency diagram obtained performing short-time Fourier transform (STFT) on radar return signal. Moreover, frequently it possible extract useful information can also task from spectrogram target. main aim paper comparing different ways exploit drone’s stages multifunctional radar. Three approaches compared: classic spectrogram-based classification; spectrum-based signal target picked after moving detector (MTD); features-based classification, undergoes detection step MTD, discriminating features extracted as input classifier. To compare three approaches, theoretical model for types drone aerial developed, validated comparison real recorded data, simulate targets. Results show third approach (features-based) not only better performance than others but one requires less modification processing power modern because reuses facility present.