6/4/2021 Allie Arp, CSL
Three CSL researchers are on a once-in-a-lifetime mission. Designing the software for a robot that will be sent to find extraterrestrial life in the frozen oceans of the moons of Saturn and Jupiter.
Written by Allie Arp, CSL
Are the inhabitants of Earth alone in the universe or is there life on other planets? This question has intrigued scientists and sci-fi fans alike for decades.
The Mars rover Perseverance is the latest attempt to analyze other planets for signs of life, but NASA believes the answers may lie even further out in the solar system. In a new project funded by the space exploration agency, three CSL professors will work to build the software for a robot capable of analyzing the oceans of a distant planet.
But why would NASA want to research the oceans of other planets when there is so much unknown about the oceans here on Earth?
“They are hoping something lives in the oceans,” said Melkior Ornik, principal investigator on the project and CSL assistant professor. “The oceans are the best chance for extraterrestrial life to exist.”
Part of NASA’s Concepts for Ocean worlds Life Detection Technology (COLDTech) program, the DRILLAWAY: aDaptive, ResIlient Learning-enabLed oceAn World AutonomY project will develop technology for a robotic lander to an ice-based moon of Jupiter or Saturn. The robot would plant itself in a particular spot, drill into the ice or scoop pieces of ice, analyze the samples, and then report the results back to researchers on Earth. There are many challenges the team will face, the least of which is figuring out how to communicate with a robot nearly a billion miles away.
“When we’re talking about this mission into the outer planets of the solar system, all communication opportunities might be scarce, damaged, and have a limited bandwidth,” said Ornik, assistant professor in aerospace engineering. “This means you can only communicate a very small amount of instructions to the rover. What we’re trying to develop is a software or tech platform for this higher level of autonomy that doesn’t currently exist in space operations.”
The lack of communication means decisions like “drill in this area” or “use this amount of force” would be made by the robot, rather than commanded by NASA. When something inevitably goes wrong, such as a piece of equipment breaking during landing, communication being temporarily lost, or the robot losing control of a certain function, the mission must continue as much as possible because of the time and financial investment that has been made. The robot must be able to overcome these obstacles, in addition to the environmental obstacles imposed by another planet, without specific direction from NASA. This means it must have resilience, in addition to autonomy.
Addressing the communications difficulties and the resilience challenges on top of building a robot capable of reaching distant moons is no easy task. When Ornik started planning this project he knew he needed additional expertise, so he turned to fellow CSL faculty members Kris Hauser and Saurabh Gupta.
“This case is a perfect example the idea behind CSL. Not one person can pull off a whole big project because no single person has all the expertise,” said Ornik. “I come from a pretty theoretical background so I’ve been working on autonomy but on a mathematical level. I don’t have a physical lab with robots or experience at implementing algorithms on complex hardware or sensors. That’s where Kris and Saurabh came in.”
Hauser brings extensive robotics experience to the team, with expertise in motion planning and system integration. The group plans to implement and test their ideas on the Robosimian, a robot already in Hauser’s lab.
“We will build physical testbeds with different terrain types and shapes for the Robosimian to manipulate,” said Hauser, associate professor in computer science. “Subjecting the robot to a variety of conditions lets us evaluate how our algorithms autonomously adapt to the unexpected.”
Once the robot lands on the surface of a moon, it is officially in the unknown. Researchers don’t know what is there, what the robot may encounter, or what environment it will be dealing with. What they do know is that the robot must perform in this unknown environment. An important part of that understanding is perceiving the environment through cameras and sensors, which is Gupta’s expertise.
“We will build techniques to adapt perception models to work well in these novel and unknown domains,” said Gupta, assistant professor in electrical and computer engineering. “As there is no source of direct supervision, we will focus on self-supervised techniques and derive supervision through consistency among multiple views and modalities.”
The team will begin their ambitious work this summer by applying their knowledge, technology, and software onto two platforms. One of the platforms is in Hauser’s lab. The other platform is an open-source simulation software platform of an imaginary world that NASA built. After they have developed the software, they will test it on NASA’s hardware platform, the Ocean Worlds Lander Autonomy Testbed.
Working on a project with the goal of finding extraterrestrial life on the oceans of distant planets is something many scientists only dream of, and the incredible opportunity is not lost on the team.
“It’s not a common opportunity to get a chance to work on one of the starting steps of a very exciting NASA mission,” said Ornik. “This mission hopes to go to places we’ve never been before and places we can only dream of, so all three of us are very excited about it.”