Nonlinear Receding-Horizon Nash Differential Game for Multi-Agent Adversarial Drone Racing

Abstract

This research presents a control method for adversarial drone racing using a receding-horizon Nash differential game. This approach realizes autonomous racing of multiple drones against each other by enabling strategic maneuvers like overtaking and blocking. The core of this method is an objective function designed to facilitate competitive interactions. To validate its effectiveness, simulations are conducted with three drones, each using a different control strategy. The results demonstrate the superiority of the Nash equilibrium strategy over both a cooperative strategy and a standard nonlinear model predictive control approach, particularly in overtaking and blocking behaviors.