Using non-cooperative games to coordinate communications UAVs

Abstract

Swarms of communications-equipped UAVs offer a new approach to the provision of temporary wide-area communications for mobile radio systems. It is clear that coverage depends on the location of the UAVs in relation to the mobiles. Dynamically planning the location of UAVs in response to the movement of the mobiles can optimize the coverage of the network. This paper introduces the use of non-cooperative games as a technique for optimizing the location of UAVs. The game-theoretic approach requires that each UAV attempts to maximize its support of the mobiles, mindful of the activities of the other UAVs. This is analogous to several companies competing for their share in a limited market. The use of game theory recasts an optimization problem as a decision problem. The method described in this paper allows the UAVs to independently reach identical decisions about their future locations without the need for a central planning agent. Exemplar results are presented for a network of three UAVs. These are used to illustrate the behaviours that the UAVs exhibit during a game.