The goal of the proposed work is to develop a decision support tool to improve shipping operations in ice covered waters. In particular, the tool will support tactical navigation decisions to optimize on a time horizon of a couple of hours to a few days.
The tool can be used to support decision making across a range of levels of ship autonomy, from providing on-demand routing advice to a bridge crew on board, to serving as a primary source of tactical decisions that the bridge crew can implement, whether they are on board or in a remote control centre, to acting autonomously to navigate an optimal route from departure to destination. Reinforcement learning, an unsupervised machine learning method, is well suited to complex problems that have competing goals, such as shipping operations.
The proposed work is well aligned with OFI’s overarching theme of Safe and Sustainable Development of the Ocean Frontier and with NRC’s Intelligent Marine Assets vision.