An autonomous seafaring project spearheaded by ProMare and the University of Plymouth, along with IBM, has been launched to commemorate the fourth centenary of the Mayflower. It could help explore the potential for machine to machine collaboration with limited communication across the oceans and even extra-planetary research.

The fully autonomous ship will sail across the Atlantic just as the original Mayflower did, which transported the first of the English Puritans from Plymouth to what would eventually become Provincetown Harbor, Cape Cod. The human-free vessel is to be packed with sensors and will collect data during its initial voyage on microplastics, as well as hopefully advancing the cause of remote research with autonomous vehicles across the oceans.

IBM
IBM

Brett Phaneuf, a founding board member of the ProMare marine research organisation and co-director of the Mayflower Autonomous Ship project, tells Techworld that designing a vehicle that does not have to accommodate the needs of humans brings a certain sense of liberation to the process. Ships must factor in the needs and safety of their human crew first and foremost, and this means providing facilities such as living quarters, toilets, and so on.

The design team instead has drawn up a sleek ship that, instead of people, will house a computer sporting IBM architecture. IBM's systems strategy leader Eric Aquaronne explained to Techworld that first of all, the organisations had to collaborate on the watercraft's image recognition capabilities, which run on IBM's PowerAI Vision technology, along with IBM Power Systems servers.

This model was trained with data and images from the Plymouth Sound bay in the English channel and, using automated identification systems (AIS), RADAR, and LIDAR for situational awareness, should enable the ship to navigate across the ocean with little guidance.

IBM's Operational Decision Manager software, meanwhile, will aid the craft in its decision whether or not to change its course – or to inform the ship the best speed and path with which to proceed. It will also be equipped with multiple camera systems and satellite communications of several types, but mostly low-bandwidth, low-orbit systems, such as Iridium.

Phaneuf said that he had already been in discussion on how to celebrate the 400th anniversary of the sailing of the Mayflower in 2020. An idea was floated to build a replica ship, but one had already been built in the late 1950s, now retrofitted and ready to sail again by this year's end.

"Instead of building something from the 17th century, why don't we design and build an autonomous ship that speaks for the next 400 years of maritime enterprise, technology, and voyages of exploration, to invoke the spirit of the pilgrims that embarked on a new beginning," said Phaneuf. "They thought it was a fine idea, and said: go ahead smart guy, do it."

Automation

Automated ships do exist to a degree, but according to Phaneuf this vehicle differs in that it is designed to be deployed for "extremely long periods of time". It is equipped with a hybrid propulsion engine that mixes sustainable energy technologies such as solar and wind with its diesel engine, meaning it could, in theory (and without unexpected incidents at sea) be deployed indefinitely.

"Opposed to building a robotic vehicle system that follows rote commands at the time it begins its operations that it follows for a relatively short period of time," said Phaneuf. The team intends to imbue its ship with the kind of intelligence so that it can operate for "extremely long periods of time" with "little or no human intervention".

Due to the people-free design, the team opted for a Trimaran model to accommodate its low-power propulsion module, along with a long and slender hull. That needed to be stabilised, said Phaneuf, so the team fitted outriggers to the Trimaran section, creating a deck that can be mounted with solar panels.

Having said that, it's certainly "not a vessel that can survive anything, and it's not intended to," said Phaneuf.

"What it's intended to do is survive across the Atlantic, and even if it doesn't, the nice part about an unmanned vessel is that nobody would be hurt. That's the goal here: to allow people to be virtually present but be safer and as a result, be more focused on the task of science, as opposed to all the other things that have to be done to keep a ship floating at sea and safe for people to operate."

The technological challenges, then, centre around the automation of certain nautical processes and actions. "That technology has rapidly advanced in the last 20 years on ships in general ... most commercial ships are highly, highly automated systems already. All the technology for the basic automation, turning things on and off, operating the windsail, re-powering a computer that needs to reboot, all these things are automated now," he explained.

Opportunities

One largely as yet uncharted matter is in regulation and legislation. At the moment, there's none at all that would prevent a research-based organisation from setting autonomous ships loose in international waters, although agencies such as the International Maritime Organisation are beginning to explore the possibilities.

Phaneuf hopes that the project could, as a "forcing function", help to focus the minds of regulator agencies about what is possible with automated maritime research, what the state of the technology is, and how it needs to be regulated – or if it needs to be regulated at all – to maintain safety at sea.

"That's one big hurdle," said Phaneuf. "The other is I don't think people widely understand how much is possible with the technology that exists today: there are vast, vast opportunities for the application of automation and autonomy particularly at sea. I think we're at the bleeding edge of the renaissance in that regard."

The organisations predict that the end result could see these ships work as a kind of research platform for interested entities to conduct marine research.

And he argues it's much safer – as even relatively high speeds through the water are slow compared to land vehicles.

"The marine environment is the perfect place to do this for a variety of reasons when compared to self-driving cars or autonomy when you think about transportation for humans," he said. "The marine environment is sparsely populated by people versus any city or town, so there's far less stuff to hit. You're in a far more hostile environment, you're in the ocean, but it's a far safer environment for unmanned vehicles and for people to interoperate because it's much more sparsely populated.

The biggest threat, somewhat predictably, remains the unpredictable behaviour of manned vessels. "As we know from experience, the fact we have to have rules at sea is so people follow them," said Phaneuf. "People don't always follow them, and even if they do, accidents ensue occasionally, so we make new rules. The ship will always follow them."

Next, by testing machine to machine communication and automation in areas of extremely limited bandwidth such as the deep ocean, this kind of project could spur advances in areas that have not yet been fully explored in the applied research world. A corollary, said Phaneuf, is outer space: "How are we really going to explore other planets within our solar system? Well, we're going to do it with machines as we already do. But those machines will become more and more sophisticated and work together cooperatively, with huge latencies in communication and low bandwidth still.

"So making smart machines on earth and learning how they operate together and what the foibles are – and how to deal with limited communications – all of that pays dividends further downstream for other areas we probably can't even imagine right now."