Plans to recreate the 1620 trans-Atlantic journey of the Mayflower colony ship with a fully autonomous, crewless vessel are one step closer, as IBM begins trials of the ship’s AI “captain” in a project that could set the scene for future crewless cargo shipping.
The Mayflower Autonomous Ship (MAS) project undertaken by IBM, the University of Plymouth and marine research firm ProMare aims to create the world’s first fully-sized autonomous research vessel that will cross the Atlantic this September.
For the last two years an AI model has been trained using a million nautical images collected from open source data sets. and exercises in Plymouth’s bay, known locally as The Sound. In order to process this database, a team in Plymouth are using an IBM Power AC922 server fitted with Nvidia V100 Tensor Core GPUs.
Upon completion the ship itself will be fitted with an IBM Power System accelerated server that will be tasked with helping the AI captain act independently on the high seas. To feed the AI captain visual information it will use a mix of radar and on-board cameras. Rigged with IBM’s computer vision technology, the ship should be able to independently spot and avoid other ships, buoys and debris.
Poor Connectivity on The High Seas
During much of the MAS’ journey it will have limited bandwidth connectivity. So in order to act autonomously the vessel will rely on a series of on-board NVIDIA Jetson AGX Xavier devices for computational power. (The devices bundle NVIDIA GPUs and Arm cores). These will process the ship’s sensory data as it makes independent decisions in the event of potential collision courses with land or other vessels.
Rob High CTO for Edge Computing at IBM commented: “Edge computing is critical to making an autonomous ship like the Mayflower possible.
“The Mayflower needs to sense its environment, make smart decisions about its situation and then act on these insights in the minimum amount of time – even in the presence of intermittent connectivity, and all while keeping data secure
The autonomous shipping market is projected to grow from £69.6 ($90) billion today to £100 ($130) billion in 2030, according to Allied Market Research.
An array of firms have been working on truly autonomous ships that do not required crews such as Rolls-Royce which is working with the Finnish state-owned ferry operator Finferries to build a car ferry. The ship uses a unique auto-docking system to independently alter course when approaching quays and docks and without any human intervention it is capable of self-docking.
Removing the crew from cargo ships has a host of benefits, obviously there is savings when it comes to crew salaries, but having no crew also means that the very design of ships can be reconfigured. A crewed ship requires a number of life supporting systems in place such as the deckhouse which houses the crew and wheelhouse. Its removal would open up space for extra cargo.
Norwegian chemical and supply chain firm Yara International is aiming to launch the world’s first fully electric and autonomous container ship this year. Called the Yaya Birkeland the cargo ship will be tasked with transporting chemicals between plants a journey that currently involves a 100 diesel truck every day.
Insurers are some way from developing policies for automous ships, but with data suggesting that more than 50 percent of all accidents at sea are caused by human error, coverage may not be a huge problem in future. Regulations may be more of a sticking point, including those that require a lookout, as well as the risk of unmanned ships being boarded by pirates without any human crew in place.