Oregon State University researchers will lead a team of scientists and engineers on a three-year, $1.5 million project to develop and test a team of robots that could travel under ice shelves and collect critical measurements about the extent of ice cavities and surrounding ocean properties.
The effort, funded by the U.S. National Science Foundation's Office of Polar Programs, is designed to help advance underwater exploration in confined and hard-to-reach environments such as cavities under ice shelves, said Jessica Garwood, an assistant professor in Oregon State's College of Earth, Ocean, and Atmospheric Sciences and the project's principal investigator.
Warming ocean conditions are causing polar ice sheets and ice shelves, which are floating extensions of ice sheets, to melt rapidly and contribute to global sea level rise, but studying the impact of this phenomenon poses a significant challenge for researchers who have limited tools to physically reach dangerous and deep, distant cavities beneath ice using existing tools, she said.
"Working in the water under ice is particularly challenging because communications are limited and there is no surface access to recover equipment," Garwood said. "We need robots that can travel into these areas and also travel back out."
The research team's goal is to develop a system with a large "mothership" robot that will carry and deploy a swarm of smaller passenger robots that could spread out and explore the waters under a melting ice shelf or other hard-to-access locations. The robots would operate autonomously and be programmed with decision-making ability based on conditions.
The "proof-of-concept" project includes building the deployment and recovery system for the parent robot and the swarm of passenger robots; developing hardware and protocols for communication and localization underwater; and navigation and decision-making algorithms that would allow the robots to adapt their behavior and data collection efforts based on the conditions they encounter.
"Once the robots are deployed, they are on their own out there. They won't be able to surface to send information, they will only be able to communicate with each other," said Garwood, an oceanographer whose past work has involved putting robots in the ocean to investigate how ocean currents move small marine organisms. "So the robots might be programmed to identify a fresh water signal coming from a melting glacier and follow that signal, for example."
During the three-year project, the researchers plan to conduct a series of tests in water, including a frozen lake in Oregon.
"The immediate goal is to develop these tools and systems," Garwood said. "The end goal is to get under ice shelves so we can investigate ice-ocean dynamics and monitor changes in ocean conditions. Such a system may also be effective in other environments, such as in the coastal ocean, where teams of resident robots could monitor ocean conditions and adapt their sampling behaviors to respond to specific subsurface signals, such as low oxygen waters."
Multi-robot systems already exist for aerial and ground environments, said Geoff Hollinger, an associate professor who operates the Robotic Decision-Making Laboratory in Oregon State's College of Engineering and is co-principal investigator on the project.
"Existing systems cannot overcome the communication, sensing, and coordination challenges imposed by the under-ice environment," he said. "Solving these problems and deploying in new environments has enormous potential to teach us about glaciers and the ocean."
The research team also includes Phil Lundrigan of Brigham Young University; Atsuhiro Muto of Temple University; Nicholas Rypkema of Woods Hole Oceanographic Institution; Yu She of Purdue University; and Xi Yu of West Virginia University.
The project resulted from a National Science Foundation-sponsored Ideas Lab workshop that brought together scientists from several disciplines, including robotics, polar science, oceanography and engineering, to brainstorm innovative solutions to advance underwater science.
