C3R: A novel Classification model-based Coordination method for online Conflict Resolution of Multiple Unmanned Aerial Vehicles

The wide applications of Unmanned Aerial Vehicles (UAVs) in the low altitude urban airspace require the elaborate airspace conflict resolution mechanism. The online decentralized coordination method is the essential part of this mechanism. This paper proposes a novel classification model-based coordination method. Firstly, the collision cone-based safe separation constraints are studied and the decentralized conflict resolution problem is analyzed. Secondly, the right turn policy is proposed, that to establish the basic consensus of UAVs with regard to conflict resolution. Thirdly, the linearization conflict avoidance constraint is studied. The start angle that integrates the motion direction of each given UAV and the safe separation constraint of the relevant pairwise conflict is defined. The pairwise conflicts are classified as ordinary conflicts and singular conflicts according to the start angles of related UAVs. A comprehensive coordination method is proposed based on right turn policy and the classification model. We validate our approach on representative scenarios, and the results demonstrate that our novel method is beneficial in reducing the impact on the air traffic, and is capable of dealing with complex conflict scenarios.