Air–Ground Cooperative Multitarget Hierarchical Tracking Method Based on Aerial Fisheye View

The multitarget tracking of multirobot systems has a wide range of applications, including urban security, search, and rescue. However, these tasks often take place in obstacle-rich environments, posing significant challenges, such as an unknown global map, unclear target locations, and inefficient resource allocation. In this article, we propose an air–ground cooperative hierarchical method for multitarget tracking. The uncrewed aerial vehicle (UAV) is equipped with a fisheye camera to provide a wide field of view (FOV) for multitarget detection and an initial global map generated from segmented aerial images. Multiple targets are assigned to uncrewed ground vehicles (UGVs) using the task allocation method, which is based on the pigeon-inspired optimization and auction mechanism. UGVs track targets through motion planning to avoid obstacles and keep their assigned targets in the FOV. Experimental validation confirmed the effectiveness of our method, demonstrating that UGVs can autonomously and reliably track targets in obstacle-rich environments with UAV assistance.