Spatial-temporal recurrent reinforcement learning for autonomous ships

The paper proposes a spatial-temporal recurrent neural network architecture for Deep $Q$-Networks to steer an autonomous ship. The network design allows handling an arbitrary number of surrounding target ships while offering robustness to partial observability. Further, a state-of-the-art collision risk metric is proposed to enable an easier assessment of different situations by the agent. The COLREG rules of maritime traffic are explicitly considered in the design of the reward function. The final policy is validated on a custom set of newly created single-ship encounters called "Around the Clock" problems and the commonly chosen Imazu (1987) problems, which include 18 multi-ship scenarios. Additionally, the framework shows robustness when deployed simultaneously in multi-agent scenarios. The proposed network architecture is compatible with other deep reinforcement learning algorithms, including actor-critic frameworks.