Collaborative Multi-Agent Tracking based on Distributed Learning

In order to reasonably allocate the Unmanned Aerial Vehicle (UAV) resources and keep the target tracked, we propose a collaborative multi-agent tracking method based on distributed learning. Considering the interactive fusion of information between multiple UAVs, we use the Information Filter (IF) to transform the target state fusion problem into a simple algebraic superposition. Meanwhile, the UAV's movements can be planned in real time so that the UAV formation can obtain more accurate target measurement as much as possible. The simulation experiment of collaborative multi-UAV tracking proves that the distributed multi-UAV system can make reasonable actions and track targets effectively.