Autonomous AI-powered ship, Mayflower, sets sail across the

An autonomous ship fitted with software developed by IBM set sail across the Atlantic Ocean earlier this week, according to the project organisers. The Mayflower Autonomous Ship, which was completed and launched in September, is an artificial intelligence, solar-powered marine research vessel which will sail across oceans to collect environmental data. The ship's name commemorates the crossing of the original Mayflower 400 years ago.

An environmental agenda

MAS has spent the last several months in sea trials and various research missions preparing for the big journey from Plymouth, England, to Massachusetts in the US. The ship is to work with scientists and other autonomous vessels to gather information on issues like global warming, micro-plastic pollution and marine mammal conservation. 

Weather conditions were a big factor in determining when MAS set sail. The team consulted with meteorologists from IBM's The Weather Company every day to choose the ideal departure window and increase the chances of the mission’s success.

A solely AI, automation and cloud edge device-powered journey

The ship is led by a team in Plymouth and with software engineering from IBM and using radar and GPS to navigate. Six cameras attached to the mast serve as the ship's "eyes", by feeding into an AI image recognition system to help it avoid other sea vessels and oncoming ships, as well as other hazards.

There is no human captain or onboard crew, the MAS uses the power of AI and automation to cross the ocean. The ship’s AI Captain performs a similar role to a human captain, assimilating data from a number of sources and constantly assessing its route, status and mission. Cameras and computer vision systems scan the horizon and streams of meteorological data to detect potentially dangerous storms. The small, lightweight edge devices provide just enough local compute power for the ship to operate independently, without connectivity or remote control. When a connection becomes available, the systems sync with the cloud, enabling updates and data upload.

The voyage to the US is estimated to take about three weeks.