Adaptive Multi‐Feature Validation Algorithm for Hovering UAV Detection via LFMCW Radar: Leveraging Micro‐Doppler Spectral Characteristics

With the widespread use of UAVs, timely detection of non‐compliant ones is critical for security. Radar excels in long‐range all‐weather UAV detection but faces high false alarms in cluttered environments—especially for hovering rotor UAVs, whose minimal Doppler shifts are easily masked by ground clutter. To solve this, this paper proposes the micro‐Doppler spectral feature‐guided LFMCW radar detection algorithm, a micro‐Doppler spectral feature‐guided LFMCW radar algorithm for such UAVs. Its core innovation is reconstructing signal modelling to link rotor micro‐motion with radar echo spectra, plus multi‐step decision logic: extracting range‐Doppler (RD) maps from echoes, identifying potential targets via zero‐frequency‐centred peaks and verifying via three micro‐Doppler features (uniform line spacing, bounded width and stable amplitude variation). Validated with a Ku‐band LFMCW radar prototype, simulations show lower false alarms than traditional threshold methods at 30 dB SCR. Field tests in tree‐cluttered environments confirm detection of micro‐UAVs (RCS = ) within 2 km, with distance resoslution m and speed resolution m/s, proving the algorithm's robustness and practicality.