Hierarchical Reinforcement Learning for Air-to-Air Combat
This addresses the challenge of creating effective AI pilots for military defense applications, representing a competitive but incremental advance in the field.
The paper tackled the problem of developing AI for air-to-air combat by combining hierarchical architecture with maximum-entropy reinforcement learning and expert reward shaping, achieving second place in the DARPA AlphaDogfight Trials and defeating a trained F-16 pilot.
Artificial Intelligence (AI) is becoming a critical component in the defense industry, as recently demonstrated by DARPA`s AlphaDogfight Trials (ADT). ADT sought to vet the feasibility of AI algorithms capable of piloting an F-16 in simulated air-to-air combat. As a participant in ADT, Lockheed Martin`s (LM) approach combines a hierarchical architecture with maximum-entropy reinforcement learning (RL), integrates expert knowledge through reward shaping, and supports modularity of policies. This approach achieved a $2^{nd}$ place finish in the final ADT event (among eight total competitors) and defeated a graduate of the US Air Force's (USAF) F-16 Weapons Instructor Course in match play.