The ocean is a dynamic, but often hostile environment. ProMare saw an opportunity to advance scientific research and benefit maritime industries by creating a truly autonomous vessel.
ProMare is building an unmanned vessel featuring the latest autonomous technology, incorporating IBM® edge solutions, IBM Power Systems™ AC922 accelerated servers and IBM PowerAI Vision connected to the IBM Cloud™ with IBM Cloud Object Storage™.
Slashestime-to-market by years with IBM AI and edge technology and expertise
Pushesboundaries of development of autonomous solutions and oceanographic research
Enableshuge potential commercial benefits across marine and transportation industries
Business challenge story
Tackling the ocean’s mysteries
More than two thirds of our planet’s surface is covered by water and yet 95 percent of the ocean is unexplored. Satellites have mapped the oceans from above, but what lies beneath the surface is mostly unknown. Until now, many of the ocean’s mysteries have been left undisturbed because investigating them is difficult, time-consuming and expensive.
ProMare, a non-profit corporation focused on marine exploration, saw the opportunity to break new ground. When a meeting was held to discuss how the city of Plymouth should mark the 400th anniversary of the 1620 crossing of the Mayflower from there to the New World, the company proposed that it should build a Mayflower for the 21st Century.
Brett Phaneuf, Founding Board Member of ProMare and Managing Director of MSubs (wholly owned subsidiary of Submergence Group, which teamed up with ProMare on this project), explains: “Rather than replicating the route of the Mayflower with an imitation of the original vessel, we came up with the idea for the ‘Mayflower Autonomous Ship’. MAS would be an unmanned, solar-powered ship, with the endurance and reliability to operate remotely in every corner of the globe.”
By pushing the boundaries of autonomous technology and providing new platforms for oceanographic research, MAS could advance human knowledge and innovation dramatically. To turn concept into reality, ProMare needed to overcome the unique challenges of ocean exploration.
Donald Scott, Director of Engineering at Marine Ai (a subsidiary of Submergence Group, which teamed up with ProMare on this project), says: “In the middle of the ocean, communications are severely limited. Conditions can change very suddenly, and you don’t have the option to stop and power down. With MAS, we needed to go beyond the existing technology for unmanned ships, creating a vessel that isn’t just operated remotely and doesn’t simply react to the environment, but learns and adapts independently. To do this, we had to develop state-of-the-art capabilities around navigation, collision avoidance, communications and more. We began looking for the right technology vendor to come on this journey with us.”
Making waves with AI innovations
ProMare approached IBM to help make MAS a reality. Phaneuf comments: “The first thing that came to mind when I thought of MAS was: ‘we need to build Captain Watson for Mayflower’—so we knew that we had to work with IBM. The IBM team reacted to our vision with excitement and enthusiasm, and they’ve maintained that same level of engagement with us throughout the project.”
With support from IBM, ProMare began constructing the data infrastructure for MAS. Onboard sensors for radar, sonar, video, radio, global positioning system (GPS), depth and automatic identification (the automated tracking system used by vessel traffic services to monitor ship movements) will collect data from the environment continuously. Using IBM PowerAI Vision solutions, which combine IBM Power Systems AC922 accelerated servers with deep learning toolsets and frameworks, the company is creating deep learning models to recognize navigation hazards. The Power Systems AC922 servers pair POWER9 CPUs and NVIDIA Tesla V100 with NVLink GPUs to provide massive throughput capability for high performance computing, deep learning and artificial intelligence workloads.
Scott explains: “MAS must avoid hazards such as buoys. Because this is a new application for AI, we were working with very sparse data sets—no one has bothered to put together a library of buoy images, for example. We created our own data sets and applied the IBM solution’s auto labeling feature, allowing us to start training deep learning models much sooner. By fusing environment and sensor data to generate hazard charts, MAS will determine the optimal path and speed it should take.”
ProMare is also working with IBM to create machine learning models for vehicle control, route planning and optimization, analysis of results from its automatic identification system and to avoid adverse weather. Phaneuf adds: “We’re new to the world of AI, but IBM has taken a steep learning curve and flattened it out with solutions such as IBM PowerAI Vision, designed to make model building and deployment easy.”
Extending the horizons of edge computing
To overcome limitations on weight and connectivity onboard MAS, ProMare engaged IBM to implement pioneering edge technology. Rather than leave MAS solely reliant on an intermittent connection to the cloud, the company is installing edge devices that provide limited computing capacity within a small, lightweight footprint that does not require much power. These devices will collect data from sensors and store it locally. In parallel, the on-board devices will download and update models, and will use these to make inferences on the data.
When connectivity is available, the edge devices will upload data to edge nodes, located onshore at a ProMare office. At this location, the company builds models and pushes them out to MAS as required. The edge nodes are also connected to the IBM Cloud, where data is stored in IBM Cloud Object Storage.
“We’re still learning about edge computing, but working with IBM we know that we have the benefits of built-in cybersecurity,” says Phaneuf. “When we met the IBM expert in edge technology for the first time, he blew our minds by showing up with a working model.”
Satisfying a thirst for knowledge
Working with IBM, ProMare is on track to have MAS ready for the 2020 anniversary of the Mayflower crossing. The company has drawn on IBM expertise and tools to accelerate test and development cycles and overcome unforeseen complexities.
“When we kicked off this project, we were very naïve about how complicated it would be,” recalls Scott. “We keep peeling back layers of the onion to find new challenges—luckily, IBM has usually thought of them first, and already come up with solutions. We have no doubt that working with IBM has dropped years from the timeline for MAS. Specifically, IBM PowerAI Vision enables us to build larger, more accurate models and get them into production fast.”
Phaneuf adds: “In contrast to our experience of working with other large companies, IBM has been nimble, reactive and responsive at every turn. It’s a big advantage to have their best resources on tap.”
In developing MAS, ProMare is teaming up with IBM to lead the way in innovation for autonomous technology. Starting with its inaugural journey, MAS is also expected to contribute to the advancement of oceanographic research, shedding new light on the ocean’s secrets.
“At the heart of both ProMare and IBM is a desire to explore,” says Phaneuf. “The beauty of the MAS project is that we both get to do what we do best: IBM is pushing the boundaries of technology, while we are building a ship that could have a revolutionary impact on our understanding of the ocean.”
Beyond the research value, MAS will provide insights that benefit industry too. For example, ProMare is already investigating uses for the autonomous technology in logistics and supply chain management.
Phaneuf concludes: “In creating MAS, we’re bringing autonomous technology forward leaps and bounds, which will definitely produce commercial benefits on a large scale. We keep opening doors with IBM and discovering new ways that they can help us make a splash in this thrilling area.”
Established in 2001, ProMare is a non-profit corporation and public charity that promotes marine research and exploration throughout the world. Its team of experienced archaeologists and engineers execute a variety of research projects independently and in concert with academic, corporate, public, and governmental organizations and agencies that are designed to advance man’s knowledge of history and science, and the natural world.
Submergence Group and MSubs are specialists in the design, fabrication and operation of leading-edge, cost-effective submersibles and underwater systems. With locations in Plymouth, UK and Round Rock, Texas in the US, their experienced teams possess skills in mechanical and electrical engineering, composite design and manufacture, electronics and software design, and offshore operations.