Abstract
The rapid proliferation of commercial unmanned aerial vehicles (UAVs) poses growing security, safety, and privacy challenges. This paper presents a novel frequency-domain analysis methodology to extract mechanical signatures of UAVs using backscattered optical signals from drone propellers. Through both simulations and experimental validation, the feasibility of retrieving key mechanical signatures, including the propeller's rotational speed (RPM) and the number of blades, was demonstrated. These signatures are a first step towards the real-time identification of drone models and provide insights into drone's flight behavior. The methodology, tested here with small toy drones, offers promise for real-world deployment of drone monitoring systems, complementing traditional detection techniques by operating in various atmospheric conditions. Additionally, harmonic and frequency peak analysis may allow for future improvements in trajectory tracking and payload detection. This work opens new possibilities for integrating lidar-based UAV characterization into both civilian and military airspace security frameworks.