Path planning for a maritime suface ship based on Deep Reinforcement Learning and weather data

Information analysis related to a ship and its environment is required in order to make the appropriate decisions during naval missions. However, human capabilities are no longer sufficient to reliably and rapidly process the massive amount of heterogeneous data collected by a huge lot of different sensors. That is the reason why Artificial Intelligence (AI) algorithms as decision support could help operators to choose the appropriate decisions during naval missions.This article offers a decision support model able to assist operators in predicting the path of a Maritime Surface Ship (MSS) in a dynamic environment by Deep Reinforcement Learning (DRL). Path planning of MSS in a dynamic environment is still challenging. Ocean disturbances are difficult to model thinly but collisions must be avoided. Thus, we suggest considering weather data and simplified mobile and static obstacles.