Drone Automatic Recognition Using Library of Predictor Correlator Discriminant Waveform Based on Unique Resonance Features

Recent development in drone technology along with its diverse applications has come up with the need to focus on privacy threats due to drone surveillance. Drone detection and identification are state-of-art research topics which are addressed by the RADAR signal processing algorithms. The natural resonance information extracted from the object’s backscattered signal is said to have unique resonance features according to the singularity expansion method (SEM). This article investigates the nature of resonance information using the frequency domain vector fitting (VF) method. The detailed simulation analysis is conducted by considering incident aspect angle variations and a different set of composite canonical objects. The main challenge of drone discrimination is addressed in this article by using the predictor correlator waveform. The minimum value of the quantitative measure of predictor correlator number (PDN) provides the accurate discrimination of the drone. The robustness of this approach is tested by using the noisy object’s signal for the wide range of the signal-to-noise ratios (SNR). The main contribution of this article is the detailed quantitative analysis of the real-time drone models by collecting its scattered fields. Moreover, the mathematical development to get the frequency domain response is discussed by using the measurement of scattered field data. This simulated and measurement investigation was observed to give accurate discrimination using the PDN discrimination approach based on a unique set of resonance poles and residues.