In 1620, Captain Christopher Jones sailed his ship across the tricky waters of the Atlantic from Plymouth, England to the New World. With only wind and his sails to propel the craft, the original Mayflower traveled for more than two months, a trip that should have taken around two to three weeks.
Four-hundred years later, the Mayflower will set sail again.
On September 15, 2020, a new autonomous Mayflower departed Plymouth, England, and will travel to Plymouth, Mass. However, this time the ship will be led through the water by a new AI Captain.
Comparing the two captains of the Mayflower—human or human-designed artificial intelligence. (Image courtesy of IBM.)
Marine research organization Promare and IBM, the companies behind the development of the autonomous ship, are using cameras, AI, and edge computing systems to safely navigate the vessel around ships, buoys, and other ocean hazards.
For the past two years, the team has used over a million nautical images from Plymouth Sound and open-source databases to train the ship’s AI models to sense, think, and act autonomously at sea.
Using IBM’s PowerAI Vision technology and its Power Systems accelerated servers, the company has built a deep learning model capable of recognizing navigation hazards that come into view on the vessel’s video cameras. The ship also has constant situational awareness by using radar, automated identification systems, and LiDAR—the same technology used in autonomous cars.
The two giants have also used an IBM Power System AC922 fueled by IBM POWER9 CPUs and NVIDIA V100 Tensor Core GPUs as the processor, used in many AI supercomputers. The Power System AC922, the first to use PCIe interconnect, delivers approximately two times more bandwidth compared to the PCIe Generation 3 interconnect. POWER9 supports up to 5.6 times more I/O and two times more threads than its x86 contemporaries. The combination of the three systems provides massive throughput capability for high performance computing (HPC), deep learning, and AI workloads.
Fueled by a solar-fed battery, the 49-foot-long aluminum ship has a single, central tower to keep its radar, cameras and GPS antennas above water, with a diesel generator for backup power.
Traveling at about 10 knots, the ship will process data locally on multiple NVIDIA Jetson AGX Xavier devices, which will enable the ship’s AI to take the time needed to make decisions while reducing the amount of data flow and storage. Experts will use the information to update the deep learning models and push them out to the ship as required.
The craft will also use Operational Decision Manager, IBM’s rule management system, to help decide autonomously whether to change course or, in case of emergencies, speed out of the way, drawing additional power from its onboard backup generator. The system will also enable the ship to follow the International Regulations for Preventing Collisions at Sea and the International Convention for the Safety of Life at Sea.
As any captain or sailor knows, the weather is one of the key elements for a successful journey. With data from The Weather Company, the AI Captain will make navigation decisions, including updating the course of the ship, using Red Hat Enterprise Linux (RHEL), a cloud safety manager function, to review all of the AI Captain’s decisions. This will allow the team to ensure the ship as well as other vessels in its vicinity are safe.
IBM Services has created a new MAS400 portal to provide updates about the ship's location, environmental conditions, weather and research data. The portal features a seven-armed, stowaway octopus chatbot called Artie, who is powered by IBM Watson Assistant technology and provides key information to the public at MAS400.com.
The Mayflower project team includes, from left, Prof. Kevin Jones of the University of Plymouth; Rosie Lickorish of IBM; and Eric Aquaronne, who leads Systems Strategy, Cloud, AI & Security at IBM. (Image courtesy of IBM.)
The autonomous ship is finally ready to sail across the original route of the 1620 Mayflower to commemorate the 400th anniversary of its historic sail. Sailing with no human captain or onboard crew, the ship will be one of the first full-sized, fully autonomous vessels to cross the Atlantic in Spring 2021. The ship will spend the next six months in sea trials and research missions.
While the original Mayflower broke the restraints of the European social systems and led to a new democratic community, the new craft will likewise tackle the new challenges of climate change and the deteriorating health of the world’s oceans.
With human activity impacting the health of the ocean, climate and ecosystems, it has become necessary to collect oceanographic data to gain a deeper understanding of the health of oceans and marine life. The Mayflower will help experts spend more time researching the actions and interpretations of the data rather than the data collection. In turn, experts will be able to better identify and respond to long-term trends as well as gather critical data about ocean threats, including microplastic pollution, unsustainable fishing practices, habitat degradation, and the impact on important species, such as marine mammals.
The vessel is equipped with research pods and equipment to sample water for microplastics, detect and study whale and dolphin sounds, as well as map sea levels.
It also aims to provide a precedent for autonomous transportation.
The autonomous shipping market is set to grow from $90 billion today to over $130 billion by 2030. The development of the Mayflower will provide further research into commercial autonomous ships and the blueprint for a new generation of marine drones.
Once the Mayflower’s voyage comes to an end, the vessel’s performance data will be analyzed with IBM’s Watson AI to give insight into how the vessel overcame real-life obstacles. This will help build the future of AI code for autonomous ships and become a part of America’s legacy, much like the original Mayflower.