Sea change — setting a new course for ocean research

ProMare and IBM launch the Mayflower Autonomous Ship
by Michael Tucker
12-minute read

“The single biggest challenge is the ocean itself,” says Brett Phaneuf, Co-founder of ProMare, a marine research and exploration organization. “No ship has ever been built that can survive whatever the ocean could throw at it.”

In 1620, despite the vast challenge ahead, 102 passengers of the original Mayflower sailed from Plymouth, England on a hazardous voyage to the New World, inspired by hopes for a better future. And in 2016, when the Plymouth community met and discussed how to commemorate the upcoming 400th anniversary of the Mayflower’s journey, the age-old challenge of the sea and a new inspiration for discovery prompted Phaneuf to stand up and ask, “Why not use this opportunity to advance into the future, rather than reminisce about the past?”

Phaneuf had moved to the UK some years earlier to establish MSUBs, a firm that that designs and builds mini submarines, primarily for the security and defense industries. He had long been interested in autonomous systems but hadn’t yet turned his hand to an autonomous surface vessel. Now, he saw his opportunity.

“When the city of Plymouth talked about building a replica of the original Mayflower, I told them, ‘There already is one in Massachusetts — I grew up not too far from it,’” says Phaneuf. “Instead, we should speak to the next 400 years of the maritime enterprise. Let’s be inspired by what the Pilgrims did and jump off into a new beginning.”

The timing was perfect for Phaneuf to pitch his idea: an autonomous, crewless vessel that would cross the Atlantic, trace the route of the original Mayflower, and perform vital ocean research during the voyage and well into the future. “That was when the idea of the Mayflower Autonomous Ship was born,” Phaneuf says. “We now call it MAS for short.”

Driving Phaneuf’s vision for MAS is the growing need for more scientific knowledge about the ocean. “Despite two-thirds of our planet’s surface being covered in water, we’ve explored less than 5%,” he says. “That ignorance should scare us because the ocean controls the entire climate of our planet.”

Sensors, edge devices, decision automation and computer vision enable

full

autonomy (no human intervention required)

MAS's Transatlantic voyage is

currently

en route

An autonomous ship has the advantage of being purpose-built for whatever its function is. The ship becomes a machine rather than a floating hotel. “If you take the human factor out of ships, it allows you to completely reimagine the design,” Phaneuf says. “You can focus purely on the mechanics and function of the ship.”

However, in practice, Phaneuf and ProMare knew that the key to putting an autonomous ship at sea, without humans at the helm, was computing technology that could power onboard intelligence. For MAS to make the leap from vision to vessel, ProMare needed a technology partner with deep expertise in AI, edge computing and scientific research, as well as experience with the global shipping industry.

By leveraging AI and other new technologies and developing a new class of efficient, crewless and environmentally-friendly ship, MAS will start a new era of marine exploration just when we need it the most.
Brett Phaneuf
Co-founder of ProMare and Director of the Mayflower Autonomous Ship Project
The earth from space

In late 2016, when ProMare’s naval architect presented the MAS concept at a conference in Switzerland, representatives from IBM were in the audience — and were intrigued by Phaneuf’s vision.

“During the early days of the Mayflower project, I said to my CTO that I needed him to build me a ‘Captain Watson’ because I grew up watching the IBM Watson team develop ever-more powerful AI,” says Phaneuf. “That’s what we needed to take our ship across the sea. We were lucky to meet an incredibly generous group of people at IBM who had even more capability and wanted to help.”

Data we need for the ocean we want

Photo 1: Rosie Lickorish, Software Engineer and Oceanographer, IBM Research (Credit Tom Barnes for IBM/ProMare) (edited)

Covering 71% of the Earth’s surface, the ocean generates more than half of the world’s oxygen, regulates global climate and provides a heat sink to reduce the effects of global warming. However, despite the ocean’s enormity, it is not immune to human activity. Today, the ocean is more polluted, warmer, more acidic and stormier than ever.

To help reverse oceanic decline, the United Nations declared 2021 – 2030 to be the Decade of Ocean Science for Sustainable Development. This global initiative will promote a better understanding of the ocean to protect it and manage it more sustainably — and the key to understanding the ocean better is more and better data.

Nevertheless, gathering data about a system as vast and complicated as the ocean is enormously expensive. Conducting research in an environment as unforgiving as the ocean also puts ships and crews at high risk. The practical impact of this cost and risk is that vast areas of the ocean’s surface still remain unexplored. Huge gaps in knowledge persist about climate change, plastics pollution, habitat degradation, marine life conservation and other important topics.

“Most of the ocean data we do have is biased to where the commercial shipping routes are,” says Phaneuf. “To explore the rest we can’t afford, as a society, to keep building giant USD 50 million – USD 100 million research vessels. Even if we did, we’d never have enough. But at the same time, we can’t afford to ignore the growing need for ocean research.”

Rosie Lickorish, Software Engineer and Oceanographer, IBM Research (Credit Tom Barnes for IBM/ProMare) (edited)
The Launch of the Mayflower Autonomous Ship, September 2020 (Credit Tom Dickinson for IBM/ProMare) (edited)

Photo 2: The Launch of the Mayflower Autonomous Ship, September 2020 (Credit Tom Dickinson for IBM/ProMare) (edited)

Autonomous research vessels such as MAS — integrated with other shore-based, ship and satellite networks — can collect data about the ocean at a scale and cost-effectiveness far beyond what is possible with today’s relatively small fleet of crewed research vessels. “That’s why I became interested in using robotics and autonomy in ocean exploration. It opens up parts of the ocean that simply aren’t accessible for humans,” says Phaneuf.

As envisioned by ProMare and the MAS project, the future of ocean research is a hybrid blend of crewed and crewless vessels and devices, large and small, with varying degrees of autonomy — all working together to increase the quantity and quality of data.

Until recently, the goal of amassing more ocean data without sending more sailors and scientists on costly, risky voyages seemed out of reach. Everything depended on one question: Are AI, edge computing and related technologies ready and robust enough for real autonomy?

“Throughout history, ships have been at the absolute edge of technological development because they were required for commerce and moving people around,” says Phaneuf. “And it’s no surprise that with MAS, that’s happening on ships. They’ve always been at the forefront.”

Aye, aye AI Captain

Photo 3: Brett Phaneuf, Co-Founder of ProMare and Director of the Mayflower Autonomous Ship (Credit Tom Barnes for IBM/ProMare) (edited)

With its sleek, Star Trek-reminiscent shape, the MAS’s highly stable trimaran hull can skim over the waves at over 10 knots, driven by a solar-powered, hybrid-electric propulsion system. MAS is also fitted with precision inertial navigation, Global Navigation Satellite System (GNSS) positioning, radar sensors, satellite communications and meteorologic instrumentation.

Built in Poland to ProMare’s specifications, the 5-ton, 15-meter-long vessel incorporates many advanced marine architecture features, all designed to withstand the stresses of extended trips at sea. But there are no bunks or galleys for a captain or crew on this Mayflower. Instead, it is cognitive computing technology from IBM, already proven in other industries, that pilots this new kind of ship and helps keep it on an even keel.

“We needed to go beyond existing technology for unmanned ships to create a vessel that isn’t just operated remotely and doesn’t simply react to the environment, but learns and adapts independently,” says Don Scott, Director of Engineering at the Submergence Group, a ProMare affiliate. “When we conceived this vessel four years ago, I wasn’t sure we were going to be able to host a system like this on a ship. But recent advances, especially in edge computing, have made this possible.”

Brett Phaneuf, Co-Founder of ProMare and Director of the Mayflower Autonomous Ship (Credit Tom Barnes for IBM/ProMare) (edited)
Don Scott, Director of Engineering at Submergence Group and CTO of MarineAI (Credit Tom Barnes for IBM/ProMare) (edited)

Photo 4: Don Scott, Director of Engineering at Submergence Group and CTO of MarineAI (Credit Tom Barnes for IBM/ProMare) (edited)

The onboard “brain” that gives MAS full autonomy — the ability to operate without human intervention — is nicknamed “AI Captain” by the team. Just as any human captain requires years of training before being entrusted with a ship, the MAS’s AI Captain underwent years of training to learn the ropes.

Using inference algorithms and models generated from IBM® Visual Insights computer vision technology, the AI Captain was trained on over one million nautical images so it could recognize ships, debris, bridges, pieces of land and other hazards. IBM Power® Systems servers were used to meet the significant machine learning processing demands required for AI Captain training. Additional math modeling for decision support was provided by IBM CPLEX® Optimizer.

As soon as the AI Captain receives actionable data from on-board cameras, radar, sonar, AIS and other equipment, it draws on IBM’s automated rule management system, called IBM Operational Decision Manager (ODM) to follow International Regulations for Preventing Collisions at Sea (COLREGs) and other seagoing conventions. Widely used across the financial services industry to approve loans and personalize customer offers, ODM provides a transparent record of decision-making processes to avoid “black box” scenarios when it isn’t clear why a particular decision was made.

Photo 5: The Mayflower Autonomous Ship (Credit Oliver Dickinson for IBM/ProMare)

To sense its environment and make smart decisions independently without having to be connected to an onshore computer, MAS has 15 edge computing devices to process data onboard. Decision logic, based on ODM software, also runs as an edge workload on the ship.

All edge devices are orchestrated securely by IBM Edge Application Manager, which deploys software out to the sensors positioned at the edge of the ship. The IBM Maximo® Application Suite provides intelligent asset management and maintenance. “I’ve called MAS the ultimate edge device,” says Scott. “All the cameras are intelligent edge devices. Even the battery management system is an intelligent edge device.”

Once underway, MAS’s AI Captain relies on ODM recommendations and current and forecasted updates from The Weather Company®, an IBM Business, to continuously assess options, steer clear of hazards, make navigation decisions and perform the research mission at hand.

The Mayflower Autonomous Ship (Credit Oliver Dickinson for IBM/ProMare)
A Visualisation of MAS's AI Captain (Credit Birmingham University for IBM/ProMare)

Photo 6: A Visualisation of MAS's AI Captain (Credit Birmingham University for IBM/ProMare)

Keeping all data and applications seamlessly integrated, secure and connected from every edge of the MAS network are IBM Cloud® and IBM Cloud Object Storage, designed to support mission-critical workloads everywhere, even in the middle of the ocean.

A Safety Manager function, running on Red Hat® Enterprise Linux® (external link), also backstops AI Captain decisions to help ensure the safety of MAS and nearby vessels.

As AI Captain training continues and it logs more hours of autonomy at sea, ProMare is already thinking about the future of this technology, even beyond the MAS project. “The nice thing about the software we’re developing for the AI Captain is that it can also be deployed on manned vessels,” says Phaneuf. “It can augment the ability to execute a task and free up the creative part of a person to be a better human, and in this case, be a safer human.”

Sea legs for science and commerce

Photo 7: The Mayflower Autonomous Ship in sea trials, February 2021

Despite disruptions related to the worldwide COVID-19 pandemic, key MAS project milestones were achieved in 2020.Even as the vessel was completing construction, the ProMare and IBM team developed and tested the AI Captain at an R&D station in Plymouth. Jutting into the city’s main harbor from the end of a pier, the R&D station experienced all the action of a ship’s bridge and was crucial in teaching the AI Captain how to recognize and navigate around ships, buoys and other water hazards.

As the vessel’s main and outrigger hulls were transported from Poland to England for final assembly, more detailed work started on the principal purpose of MAS: to gather scientific knowledge about the ocean. The vessel’s modular design includes three science bays where research kits can be snapped in or out, depending on research topics. IBM Research® has developed three initial research projects for MAS: measuring sea levels and wave heights, analyzing the chemical composition of seawater, and listening to whale songs through hydrophones.

“One of the things that really excites me about this project is that we’re going to use technologies I work with every day to revolutionize the way we capture data in the ocean,” says Rosie Lickorish, an IBM researcher and software engineer who also has an ocean research background. “We’re going to use AI, autonomy, robotics and edge computing to create new sensors and devices that allow us to scale far more efficiently how we collect data from the ocean.”

The Mayflower Autonomous Ship (Credit Oliver Dickinson for IBM/ProMare)
The Mayflower Autonomous Ship at Launch, September 2020 (Credit Tom Barnes for IBM/ProMare)

Photo 8: The Mayflower Autonomous Ship at Launch, September 2020 (Credit Tom Barnes for IBM/ProMare)

The red-letter day in 2020 was September 16th, when the fully assembled MAS vessel was officially named and launched into Plymouth Harbor. Also launched was Mas400.com (external link), a live mission portal that enables the world to follow the MAS project and receive updates on its voyages.

Now that MAS is in the water, the vessel will undergo several months of sea trials and research missions, leading up to its maiden voyage from the UK to the US, scheduled for April 2021. “We’ll be using sensors aboard the ship to calibrate readings [from satellites and near-shore sensors],” says Lickorish. “That’s going to provide us with valuable information for things like shipping and ports.”

Lessons learned from MAS will not only have scientific value but will also drive practical and sustainable advances for many organizations that operate on or near the ocean, such as shipping and logistics, oil and gas exploration, and security and defense-related industries.

“While the autonomous shipping market is set to grow from USD 90 billion today to over USD 130 billion by 2030, many of today’s autonomous ships are really just automated — robots which do not dynamically adapt to new situations and rely heavily on operator override,” says Scott. “Using an integrated set of IBM’s AI, cloud and edge technologies, we are aiming to give the Mayflower the ability to operate independently in some of the most challenging circumstances on the planet.”

While the countdown continues to the start of MAS’s 3,220-mile UK-to-US voyage in the spring of 2021, ProMare has plans to go further. “We’re going to go to the Arctic and we’re thinking about global circumnavigation,” says Phaneuf. “We’ll continue to use MAS as a platform for many years into the future, and hopefully, she will inspire the next generation of ships as well.”

As April 2021 draws closer, Phaneuf and his team think about the original Mayflower crossing and share many of the same hopes and fears the Pilgrims felt when they embarked into the unknown. “I think what we have most in common with them is we’re not sure we’re going to make it either. But we’re willing to take the risk,” says Phaneuf. “I focus on getting the resources we need with the help of our team, and IBM in particular. I don’t think I will relax at all until MAS crosses the ocean and parks in Plymouth, Massachusetts.”

ProMare logo

About ProMare

Established in 2001 to promote marine research and exploration, ProMareExternal Link is a nonprofit corporation and a 501(c)(3) charity based in Chester, Connecticut. Its team of experienced archaeologists and marine professionals execute research projects independently and with academic, corporate, public and governmental organizations to advance knowledge of ocean-related history and science.

Solution components
ProMare logo

About ProMare

Established in 2001 to promote marine research and exploration, ProMareExternal Link is a nonprofit corporation and a 501(c)(3) charity based in Chester, Connecticut. Its team of experienced archaeologists and marine professionals execute research projects independently and with academic, corporate, public and governmental organizations to advance knowledge of ocean-related history and science.

Solution components