Unmanned Aerial Vehicle Recognition Based on Clustering by Communication with Local Agents and Multiple Polarimetric Features

Radar Polarimetric information can effectively expand the understanding of target shape and structure, and have potential of exploring materials and local fine features of targets. In this paper, based on the theory of coherent and incoherent polarimetric decomposition, we extract features of multiple types of unmanned aerial vehicles with full polarization. Then communication with local agents (CLA) method is used to overcome shortcomings of difficulty of obtaining target training samples under real electromagnetic environment conditions. Based on the extracted polarimetric scattering mechanism, we can realize the accurate identification of UAVs without training samples. The validity of the proposed algorithm is verified based on electromagnetic data measured in anechoic chamber and calculated by the software. The experiment proves that the recognition rate of the proposed algorithm is higher than that of conventional algorithms, and it has the ability to characterize the fine structure of unmanned aerial vehicles.