Autonomous Drone Combat: A Multi-Agent Reinforcement Learning Approach

Drones have become essential tool in various industries, from agriculture to surveillance and are now increasingly deployed on battlefields for detection, recognition, identification, and combat. While most systems remain controlled by human, the shift toward autonomy is intensifying, driven by breakthroughs in artificial intelligence, notably in reinforcement learning and scalable simulation techniques. This paper presents two contributions. A multi agent reinforcement learning environment for drone combat built on IsaacLab. An in-depth comparison between decentralized learning and self-play scheme in competitive settings. Our work confirmed the benefits of self-play methods for autonomous drone combat.