400 Years Later Mayflower Is Preparing To Make Autonomous AI Voyage
In 1620, The Mayflower Set Sail On Its Journey From Plymouth, England, To The U.S., Carrying The First Puritans Across The Atlantic…
to colonize the New World.
Now, 400 years later, another Mayflower is preparing to make the same voyage. The big difference? This time there aren’t any passengers, there’s no captain, or even a crew. Instead, almost everything is powered by artificial intelligence (AI)
Nearly every move of IBM’s next-generation Mayflower Autonomous Ship (MAS) is regulated by AI without any need for physical human assistance, from its onboard science experiments to charting the 5,000-kilometer route across the ocean.
Not only is the ship unmanned, it also runs on renewable solar power energy – which means it can stay at sea unattended for months at a time with little negative impact on the environment.
And according to IBM’s chief technical officer for England and Ireland, Andy Stanford-Clark, the design could be a “tipping point” for the marine industry and the future of shipping.
Mayflower Autonomous Ship factsheet
- Size: 15 meters long, 6.2 meters wide
- Weight: 5 tons
- Max speed: 10 knots
- Mission: to collect oceanographic data
- Power: solar-driven hybrid electric motor
- AI cameras: 6
- Sensors: 30+
- Software: IBM Visual Insights computer vision technology, IBM Operational Decision Manager automation software, data from The Weather Company
- Designers: ProMare, IBM and a global consortium of partners
- Octopus aboard: 1
- Crew needed: 0
After two years designing the ship and programming its cutting-edge technology, the new-generation Mayflower was unveiled on 16 September, the 400th anniversary of its namesake’s maiden voyage.
Stanford-Clark, one of the ship’s chief architects, tells CGTN Europe the city of Plymouth had originally come up with a more traditional approach to celebrate the date – building a historical replica of the original Mayflower.
“But other voices argued that we should look forward to the next 400 years,” he explains, by building a ship that incorporated “gathering data, artificial intelligence and all the technologies of today.
“And that’s how the idea for the Mayflower Autonomous Ship came about.”
Made in partnership with marine research charity Promare, the University of Plymouth and autonomous underwater vessel expert MSubs, the next-generation Mayflower is designed to provide research on vital issues such as global warming, pollution and marine-life conservation.
Boasting a host of experiments for recording marine data from microplastics to whale song, the ship even carries what the British scientist describes as “an electronic tongue, which uses the same technology as our tongue, to test for different tastes in the water using chemical sensors.” But what makes the ship so special is that all of this is controlled by the Mayflower’s AI captain.
How does a ship chart its own course?
The MAS’s AI system, which uses 30 sensors and six AI cameras, is hinged on the latest computing technology – automation software, computer vision technology and the company’s Red Hat Open Source software.
Stanford-Clark explains that through computer learning, the ship’s automated navigation system has been trained on thousands of images of things in the sea – “ships, seagulls, canoes, bits of wood, icebergs” – so it can recognize what’s in front of and to the side of the ship.
“That advises the AI captain to factor in things like the weather forecast, data from a radar and also the planned mission which it thinks it’s going to go on, to then plot a course to obviously avoid ships, to steer away from storms,” he says.
The Mayflower’s progress can then be followed by researchers and the public on its MAS400 interactive web portal, which Fredrik Soreide, scientific director of the project, describes as “one of the most advanced ocean mission web portals ever built.”
It gives real-time updates on the ship’s location, environmental conditions and data from its various research projects, all animated by IBM’s “stowaway” octopus chatbot called Artie.
According to Soreide, “users can even help Artie the Octopus fish out surgical masks, cigarette butts and other increasingly common forms of ocean litter from a virtual ocean of facts and data.”
But the technology isn’t only there for novelty.
No people, no problems
The big advantage of having an AI captain and no crew, according to Stanford-Clark, is that the ship “doesn’t get bored.”
“We don’t have to pay the crew. It’s not dull, dirty or dangerous for them,” he says. “You can spend months and months at sea just going up and down collecting data.”
With so much unknown about the effects of climate change and pollution, the scientist says: “Without that data, we’re just operating blind. So having this autonomous ship that can go out and gather data is exactly what we need to advance the world’s understanding.”
It is already much smaller and lighter than its predecessor – the next-generation model weighs five tons in comparison with the original’s 180, but by being entirely autonomous, the Mayflower doesn’t need to make room for beds, food and water storage, toilets or kitchens.
“All the space inside the Mayflower can be given over to the experimental pods, which are one-meter-cubed bays down the length of the ship and we can put all experiments in that,” says Stanford-Clark.
However, one of the potential issues was the AI would require a lot of computational power, and as the ship is powered by solar panels with a back-up bio-diesel generator, it has to use its energy economically.
“Normally you expect an AI to use quite a lot of power,” says Stanford-Clark, “but we do all the hard work beforehand in the Cloud.
“We train the deep learning model using huge amounts of computing resource. Then the bit we download into ship is actually quite small and can run on a low-power GPU device.”
The Mayflower was originally meant to sail to the U.S. this September, but due to the COVID-19 pandemic, the launch has been delayed to April 2021.
In the meantime, Stanford-Clark says the team will spend six months trialing the ship at sea and in UK coastal waters before its transatlantic voyage, making sure it’s used to the salty environment and that none of its wires can shake loose at sea.
“If anything breaks during the journey, well, I’m afraid it’s on its own,” he says. “Literally, when it gets into the ocean, it’s autonomous – there’s nobody there.”
However, the Mayflower’s architects are confident the boat will be shipshape and hopeful the groundbreaking experiment will have a lasting impact in their field.
“We think this is a really exciting start to a new future for the marine industry,” says Stanford-Clark. “At the moment, people go away from their families for weeks or months on end and they only get a few days of research done with their science experiment. It’s very wasteful of time, resources and money.”
He believes the Mayflower’s autonomous technology brings the industry one step closer to solving these issues and in the future, it could even be applied to cargo ships and water taxis. “We think this is a real tipping point for the marine industry,” he says.
Whether the 21st century Mayflower will have as much impact on the world as its illustrious predecessor is yet to be seen, but it will definitely reach its goal much faster.
“The original Mayflower crossing took about two months. We hope to do it somewhat faster, in about two weeks,” says Stanford-Clark.
But he’s in no hurry. “It’s not a race – we’re not going to run out of fuel or water or food, because nobody’s onboard,” he says. “We’ve got all the time in the world.”
This news was originally published at newseu.cgtn.com