Game-Theoretic Decision-Making and Payoff Design for UAV Collision Avoidance in a Three-Dimensional Airspace.

The article focuses on a novel game-theoretic decision-making framework for collision avoidance of unmanned aerial vehicles (UAVs) operating in three-dimensional (3D) airspace. It formulates UAV collision avoidance as a repeated two-player normal-form game, where the "Ego" UAV and an adversary UAV select from a set of maneuvers—keeping velocity direction, changing direction to the right, or descending altitude—with payoffs designed to balance safety and efficiency. The approach integrates rule-based cognitive information and critical time-to-collision metrics to construct payoff functions, enabling the computation of Nash equilibria that guide optimal collision avoidance actions. Simulation studies demonstrate the method's effectiveness in various multi-UAV scenarios, highlighting its computational efficiency compared to optimization-based methods and its adaptability through repeated game play. The paper suggests future work to extend the framework to more complex scenarios with additional UAVs and actions, as well as to incorporate data-driven tuning of payoff weights.