Autonomous Marine Operations Planning

The project Autonomous Marine Operations Planning is investigating and developing Artificial Intelligence and Machine Learning toolkits to provide decision support for marine operations planning, to minimise the fuel usage across marine vehicles in multiple-year science cruise planning. The underpinning route-planner toolkit of this project is leveraging a diverse range of datasets and forecasts from environmental information (e.g. IceNet), to generate fuel efficient navigational routes between waypoints. These navigational routes are then being utilised in broader marine operations to determine the optimal plan for science cruise, using information from the route-planner to determine the carbon-cost and time for each of the tasks in the plan. These toolkits will have a far-reaching applications across many research areas within NERC allowing scientist and planners the ability to determine, and minimise, carbon-cost across a range of scientific activities. Outlined below is a more in-depth outline of the key sub-projects with links to the relevant pages.

Science cruise planning is a huge task, taking a team of operations experts many months to complete. The job involves planning the activities of all vessels so that they are in the right places at the right times to collect data, rendezvous with scientists and supply the research stations. A cruise extends over several months, with a vast array of science goals being addressed over huge geographical areas. The vessels include ships and marine autonomous vehicles such as gliders, powered AUVs and aerial drones. One of the most important goals for cruise planning is to optimise the use of fuel so as to minimise the carbon footprint of the cruise. This depends on the ability to generate highly efficient navigational routes between waypoints that take account of the complex environmental conditions. The optimisation of these tasks is notoriously difficult for human operators because of the combinatorial explosion of alternative ways in which things can be done and orderings of different activities. Furthermore, the changing weather conditions make these choices impossible to contemplate. This project acts to tackle the task of cruise planning autonomously, using a combination of data-informed AI task planning and route planning, grounded by collaboration with leading experts in operation planning.

The project can be broadly separated into the sub-projects outlined below:

PolarRoute – Optimal time & eco-friendly navigational routing
AI Cruise Planning – Data-informed AI task planning for a series of tasks across a scientific cruises
IceNav – Risk aware In-Ice navigation using machine learning emulation