Rotor unmanned aerial vehicle localization in the building sheltered area based on millimetre-wave frequency-modulated continuous wave radar

Abstract

Rotor unmanned aerial vehicles (UAVs) play an important role in both military and civilian fields nowadays. The safety risks associated with the UAVs increase the urgent need for detecting UAVs in urban environments as well. Moreover, UAVs are easily located in the non-line-of-sight (NLOS) building sheltered area relative to the radar, making it very challenging to detect and localise. A novel algorithm for localising the rotor UAV in the common L-shaped street building sheltered area is proposed. First, the authors establish a multipath signal model for a rotor UAV hovering over an L-shaped street scene, leveraging the frequency-modulated continuous wave signal. Then, the multipath information of the UAV is extracted by identifying the rotating periodicity of the blade embedded in the time-frequency spectrum of the received signal. The back projection imaging is then conducted on the UAV-related multipath. After extracting multipath ghosts in the image, the street area, where the UAV locates, can be determined, and the UAV is further localised using the path reflection characteristics of this area. Simulations and practical experiments based on millimetre waves indicate that the proposed method can enable high-accuracy estimation of rotor UAV in the NLOS building sheltered area.