Bridging Simulation and Reality: Modeling Wind-Driven Micro-Doppler Variability in UAV Radar Signatures

Understanding the impact of environmental disturbances on radar micro-Doppler signatures is critical for the reliable detection, classification, and tracking of small aerial targets such as quadcopters. In real-world operations, wind-induced motion is inevitable and can distort the ideal micro-Doppler signatures typically generated through simulations, as collecting high-quality real-world radar returns is often challenging due to logistical and safety constraints. This study investigates the influence of wind-like perturbations on the radar return from a hovering quadcopter by modeling its translational velocity as an Ornstein–Uhlenbeck (OU) stochastic process—a physically plausible representation of temporally correlated wind-induced drift. Through comparative analysis of clean (ideal hovering) and wind-induced (non-ideal) cases, we observe significant deviations in spectral characteristics, including energy broadening and raised spectral floors in the power spectral density (PSD), degradation of harmonic structures in scalograms and in persistence spectrums, and increased spectral centroid and spectral spread fluctuations over time. These findings demonstrate that even moderate wind-induced drift can introduce ambiguity in signature-based drone classification systems. By providing a controlled yet realistic simulation framework for wind effects, this work offers key insights into the design of robust radar-based drone monitoring systems intended for deployment in complex, real-world environments.