Cooperation between Ground and Aerial Robotic Platforms for Target Detection and Tracking using Deep Learning

This work researches the collaboration between unmanned ground vehicles (UGVs) and unmanned aerial vehicles (UAVs) to detect and track moving targets on the ground. The research includes three main components: a ground control station (GCS), a UAV, and a set of UGVs. The UAV provides a comprehensive view of the terrain for target detection. The UGVs are responsible for directly pursuing the target, receiving a set of routes. The GCS is responsible for processing aerial images and coordinating the UGVs. GCS uses deep learning to segment aerial images semantically, identifying ground structures inaccessible to UGVs. As a result, a navigability map of the terrain is produced. The GCS then uses this map to control the UGVs, estimating and transmitting possible routes to the desired target.