Bye Bye, Captain? The Sooner-Than-Expected Return of a Crewless Sea Voyager

On 18th June 2021, researchers had to call a robotic trimaran back to shore that was on its way to make history. The AI-Driven MAS 400 or Mayflower Autonomous Ship 400 is the largest autonomous sea voyager of its kind that is built to navigate an entire ocean without the support of a manned captain and crew. The ship was called back for an inspection when it was just 350 miles off its departing shore of Plymouth, England. It had completed only 10% of its total journey.  The MAS 400 is a joint venture by the non-profit ocean research organization ProMare and tech giant IBM in a scientific endeavor to collect data related to climate change, pollution and conservation of marine life.   

mas400
Courtesy: IBM Image Gallery, Mayflower400UK.org

The Mayflower Autonomous Ship 400 (MAS 400) departed for a historic trip from Plymouth, England to Massachusetts, USA on Tuesday, 15 June 2021, marking the 401th anniversary of the 1620 transatlantic voyage of its namesake. In 1602, a sailing ship named Mayflower undertook a historic voyage across the Atlantic ocean carrying 102 passengers to colonize the region of present day Cape Cod, Massachusetts. The futuristic version of the ship is powered by solar panels, steered by radar technology and led by an AI-powered captain in charge of making real time decisions. 

About the MAS 400

Less than 3 percent of the weight of the original Mayflower, the MAS 400 is led by an intelligent bespoke AI Captain built by MarineAI and uses the functionality of the IBM Visual Insights Computer Vision Software and IBM Operational Design Manager Automation Software. The data used by the captain is gathered from 30 sensors including radars, GPS, nautical charts, altitude sensors and fathometer (also known as echo sounder or sonar).  

Since the ship would not always have connectivity, it also uses 15 edge devices using IBM Edge Computing Software to process data locally rather than sending it to a centralized data repository. To classify and maneuver around buoys, barriers or debris in its way, the MAS’s AI Models have been trained for years using over a million nautical images captured from cameras of Plymouth Sound, UK and other open source databases. 

mas400
Courtesy: IBM Image Gallery

What went wrong?

Despite the training and careful preparation, the AI-powered captain seemed to have missed something when the ship was a few hundred miles into its journey. It was going too slow, most probably to save power. The researchers back in the control station noticed the reduced speed and realized the absence of any rationale by the AI Captain behind it. What was more concerning was that the AI Captain had not raised a red flag about a malfunction that might have caused the ship to move at half the optimal speed. 

Additionally, it was impossible to figure out what was wrong as the cameras monitoring the components of the ship didn’t cover everything. If, in case, there was a mechanical fault that was wasting power, the crew-less ship will not have been able to steer through the upcoming turbulence of the North Atlantic Gulf Stream. So to avoid the chances of any mishap, the researchers thought it best to call the ship back home and see what was wrong for themselves. The fault still hasn’t been found as the researchers await the arrival of the MAS 400 in a few days. 

What’s next? 

ProMare is ready to inspect the MAS 400 once it reaches back to base and then send it off on its destined transatlantic journey. It is expected that, if complete, the ship will take only about three weeks to cover the 3200 mile distance from England to Massachusetts, compared to a whopping 66 days taken by its legacy ancestor in 1620. In light of this challenge faced by MAS 400, a statement by ProMare co-founder Brett Phaneuf holds relevance. Sharing his views on the roadblocks that MAS 400 might face, before the departure Phaneuf had commented, “The single biggest challenge is the ocean itself. No ship has ever been built that can survive what the ocean could throw at it.”

As AI continues to improve, automakers, tech giants and researchers are figuring out how to operate vehicles autonomously, whether on land or at sea. This event has allowed for scientific as well as engineering, exploratory and industrial lessons to be learnt and applied to the field of autonomous vehiclesin the future. The MAS 400 is the latest one in the line of autonomous ocean vehicles created in hopes to improve sea travel and possibly make it more and more human-free.

A.I. aboard!