Development of Active Decoy Guidance Policy by Utilising Multi-Agent Reinforcement Learning

In this paper, an AI-driven algorithm is applied to design a decoy deployment strategy which aims to increase the miss distance between the naval target and missile threat. In this scenario, three decoys are deployed from the target ship, and each decoy owns an on board jamming system which is utilized to create an artificial scatter point. Then, an Equivalent Scattering Centre (ESC) point is formed along the target-to-decoy line which represents the radar cross-section of the group of decoys and target. The decoys are trained by utilising Multi-Agent Deep Deterministic Policy Gradient algorithm for cooperative guidance which aims to move the ESC point as far as possible from the target to increase the miss distance. Preliminary results show that the proposed approach is promising to increase the survival probability of a naval platform against an anti-ship missile that is equipped with radar seeker.