Robust Adaptive OS-CFAR Detector for Maneuvering Group Target Based on Formation Prior Knowledge

Detecting the individuals such like the bird flocks or a UAV swarm is a great challenge for radar systems. The lack of prior knowledge about the characteristics of dynamic group targets leads to the performance degradation of conventional detectors. In this study, we present a robust adaptive OS-CFAR detector for maneuvering group targets, leveraging prior knowledge of their formation. The group target's state is updated and predicted using the random matrix approach. The projection length of the group target in the radar beam direction is computed using the predicted information during the group target tracking process. Using the projection length, the number of reference units occupied by the group target within the reference window is determined, and this prior knowledge is applied to update the detection parameters of the OS-CFAR detector. The simulation result shows that the algorithm can update the detection parameters timely and avoid target missed detection.