Layer Structured Multiple-UAV System Mixing Centralized and Distributed Agents

Abstract

In the control of multiple unmanned aerial vehicles (UAVs), the distributed control method has been gathered a lot of attention because of the advantage that a load of communication and calculation can be reduced. In the distributed control, the controlled object (agent) is operates depending on the local information which is acquired by communication or measurement. In general, its performance is low compared with the centralized control method. Therefore, in order to solve this problem, we propose a layered structure comprised of centralized and distributed control methods that complements the advantages of each method. The proposed structure is applied to a disaster site survey with multiple UAVs. In the investigation by UAVs, the resolution of the map information depends on the flight altitude. Based on this feature, the agents that fly at a high altitude are controlled by a centralized way and capture wide-area information with low resolution. The other group of agents fly at a low altitude, move distributed way and obtain complemented information. By designing the collaborative motion among centralized and distributed controlled agents, the proposed structure possesses both advantages of centralized and distributed control methods. The efficacy of the proposed method is demonstrated via numerical simulations.