Unique Resonance Features Based Composite Drone Recognition Using Vector Fitting Method

The natural resonance information is one of the unique features of object-backscattered electromagnetic signals as stated in the singularity expansion theory, which is utilized for object recognition. This paper focuses on a frequency domain-based vector fitting (VF) algorithm to achieve accurate extraction of resonance features. The simulation and analytically obtained resonant poles of the sphere are studied to verify the accuracy of the VF method. The recent development and demand for unmanned aerial vehicles such as drones have come up with challenges of privacy damage and the need for accurate recognition. This paper investigates the drone identification challenge by analyzing its unique resonance features. Moreover, the detailed measurement analysis is conducted to check the perfect reconstruction of the original object response by observing the percentage root mean square error. This paper contributes towards a reasonable understanding of the characteristics of the VF algorithm by creating the unique drone feature dataset.