Grouping of the UAV Swarm Based on Automatic Fuzzy Clustering

The confrontation between UAV swarms will be an important combat style in the future, and a reasonable division of such a target air fleet is instructive to the deployment of our own resources. In this paper, we combine the genetic fuzzy clustering and the cluster validation to achieve autonomous grouping of the UAV swarm. Firstly, the flight state parameters such as position, speed, and yaw angle are used as characteristic components to measure the differences among UAVs. Then, the proposed approach establishes a mathematical model for grouping of the UAV swarm by fuzzy C-means algorithm. In addition, we compare five advanced clustering validity indexes and a clustering validation method based on graph theory, from which an evaluation criterion more applicable to grouping of the UAV swarm is selected. Finally, numerical simulations are performed to verify that the proposed approach can divide the air fleet autonomously and rationally in complex task scenarios.