MARIN has kicked off a pilot project of its Data Science and Artificial Intelligence team on machine learning: AI-Sail. The challenge: Can the engineers learn a computer/robot how to sail an Optimist, using data driven and machine learning techniques?

The idea for AI-Sail is based on the observation that children intuitively can learn how to sail in an Optimist (a small, single-handed sailing dinghy intended for use by young people up to the age of 15), without understanding the details of aerodynamics and hydrodynamics. The question now is if a computer can do the same.

The pilot project resulted in a unique sight at MARIN (Maritime Research Institute Netherlands) this week, says Bas Buchner, President of MARIN, on MARIN’s LinkedIn page: a first try-out of sailing with an Optimist in MARIN’s Offshore Basin. Children of a MARIN employee sailed the boat around the Basin (watch a video here, the picture is a still taken from this video).

Also read: Looking to test your idea or concept for free at MARIN?

Modification for unmanned operation

As part of AI-Sail, MARIN will now start to modify the Optimist for unmanned performance and develop the reinforcement learning techniques to make it sail. In April next year, the real experiments are planned.

The overall objective of AI-Sail is to explore the possibilities (and limitations) of AI and machine learning for maritime purposes.

Buchner: ‘Some maritime people are pessimistic about the added value of data driven methods in our field. This project is to show what maritime potential AI and machine learning have. And as a former colleague always said: “It is better that the optimist is wrong, than that the pessimist is right!”’

Also read: VIDEO: How MARIN validates calculations of the stern wave of a mega cruise ship