Micro-Doppler radar signature identification within wind turbine clutter based on short-CPI airborne radar observations

An application of machine intelligence technique for the identification of micro-Doppler features from an airborne pulsed-Doppler radar sensor is developed. The key challenges for surveillance mode are the dynamic nature of the wind farm clutters, short-CPI length, and lack of prior information on the specific wind turbine (WT) in the site. The micro-Doppler spectrum segments based on short CPIs are used as the fundamental feature vectors for detection and classification. Both supervised and unsupervised approaches, including artificial neural network and random forest, are applied to airborne plan position indicator scan outputs. A simulator for airborne pulsed-Doppler radar operation over wind farm is used with realistic WT scattering signatures, platform motion impacts as well as the terrain clutter impacts. Based on the clutter identification result, the feasibility of detecting small moving targets in the presence of WT clutter is discussed.