A Non-cooperative Game to Coordinate the Coverage of Two Communications UAVs

Abstract

This paper considers the optimal placement of two UAVs that are providing backbone communications to a community of ground mobiles. The continuous movement of the ground mobiles motivates the UAVs to periodically relocate to maximize their joint coverage. The optimum locations for the two UAVs are decided by the outcome of a non-cooperative game. In this game the UAVs are the players, the set of possible locations are their strategies, and the number of ground mobiles uniquely covered by each UAV are their payoffs. Using game theory implies that the UAVs compete for coverage of mobiles in a similar manner to two companies competing for market share. The UAVs are power limited. Some of this power needs to be assigned to an inter-UAV link, placing constraints on the power available to support mobiles. Game theory permits decisions about navigation and radio resource to be made on-board the UAV, removing the need for a central planning function and reducing the associated communications traffic. The move towards on-board decision making allows the UAVs to act with greater autonomy where this is appropriate, for example if the control links are unavailable. This paper compares the results of the two-UAV game to a benchmark of a single UAV that can assign all available power to supporting mobiles. Initial results indicate that the two-UAV game performs favourably against the benchmark. The paper starts with a short introduction to game theory. It describes how the non-cooperative game is constructed and shows the results of simulations. It uses these results to identify the desirable characteristics for a two-UAV system and comments on potential improvements to the game.