Commercial shipping faces a perfect storm of commercial, regulatory, and market pressures to reduce its reliance on fossil-based fuels. A relatively low-cost option is to modify ship routes to take advantage of surface ocean currents to save fuel and cut emissions, without changing transit time or modifying the vessel. However, existing route optimisation technologies are only as accurate as the ocean current forecasts that they use.

In this talk, I will describe a novel approach that brings together AI-driven ocean forecasting and real-time in situ sensors to find robust, adaptive ship routes through the ocean. Using Generative AI techniques, the model can replicate a 24-hour forecast along a ship’s projected track in milliseconds, enabling rapid generation of an ensemble of hundreds of forecasts in parallel. The route optimisation system then efficiently searches for ship routes that are not just optimal (in terms of fuel consumption, CO2 emissions, safety, transit time etc) but are the most reliable route most of the time.

Forecasts and route recommendations are validated in real time via low-cost ocean and weather sensors mounted on cargo vessels. In essence, we turn ships into mobile ocean observing platforms. measuring ocean and weather in real-time, broadcast via satellite telemetry, and assimilated into the generative AI ocean forecasting system. This provides crew and fleet operations managers with the confidence that the forecast and the recommended route are the best possible and reliable under uncertain ocean conditions.

Biography:

Shane Keating is an Associate Professor of Physical Oceanography and Applied Mathematics in the School of Mathematics and Statistics at UNSW Sydney. His research uses powerful mathematical and data science tools to study our oceans from land, sea, and space.

Shane is CEO and Founder of CounterCurrent (www.countercurrent.ai), a UNSW spin-out that is using AI and satellite data to decarbonise the commercial shipping industry. Shane is also a passionate science communicator, and his popular articles about science and mathematics have been read over one million times and have been published in the BBC, the Guardian, and Time magazine.