Graduate research: The artful algorithm
“I love algorithms. I think they’re beautiful,” said Ryan Luna, who also sees beauty when NASA’s Robonaut 2 responds to an artfully written algorithm.
“My job is to develop algorithms that plan the motions of robots. We have to be able to anticipate what they will be doing in 10 minutes and in half an hour, which is an enormously complicated task,” said the fourth-year graduate student in computer science (CS) at Rice University.
In 2013, Luna was awarded a three-year NASA Space Technology Research Fellowship, and works in the Robotics Systems Technology Branch at the Johnson Space Center. His research focuses on motion planning for Robonaut 2 (R2), NASA’s humanoid robot that joined the crew of the International Space Station in February 2011.
R2, known as a “torso on a stick,” is “humanoid” because it resembles a human being, with arms, hands, torso and head, and because its mission is not the robot’s more conventional task of lifting and moving heavy objects. Rather, R2 is designed with dexterity in mind.
“He has almost the same number of joints in his hands as you have. He is designed to do the same tasks as humans do, using the same tools,” said Luna, whose doctoral adviser is Lydia Kavraki,the Noah Harding Professor of Computer Science and Bioengineering.
Luna, a native of Lovelock, Nev., earned a B.S. in computer and information engineering in 2009 and a master’s degree in computer science in 2011, both from the University of Nevada at Reno. There he studied under Konstantinos Bekris, a robot specialist who studied with Kavraki and earned his Ph.D. in CS from Rice in 2004.
R2 operates with multiple technologies, including vision systems, image recognition systems, sensor integrations, tendon hands and control algorithms. In it are almost 50 patented and patent-pending technologies, with applications in many non-aerospace fields. Since the project started in 1997, four Robonaut 2 models have been assembled, with others already in development.
“The Robonaut 2 system is a wonderful testbed for technologists like Ryan, as both an informer of hard problems to tackle and a place to integrate new technologies. Ryan’s work has been integral in creating a control paradigm for Robonaut’s climbing capabilities,” said Julia M. Badger, NASA’s Robonaut deputy project manager.
Much recent work has been devoted to Robonaut 2’s mobility. The robot’s legs, which until now have remained vestigial, will be upgraded to permit movement around the space station.
“The NASA fellowship has been very generous, giving me extreme flexibility,” Luna said. “I work on-site maybe 10 or 12 weeks a year, mostly in the summer. It’s an enormous collaborative project, with people from many different fields. I’ve gotten plenty of hands-on experience.”
While at Rice, Luna has also worked as a key developer for the Open Motion Planning Library, an open-source software package for sampling-based motion planning of high-dimensional and complex systems. In 2014, he received the Andrew Ladd Graduate Fellowship for Excellence in Computer Science.
Luna already has a job lined up after commencement, when he goes to work for Google, assigned to its self-driving car team. “I’ll be building on the experience I acquired working on the Robonaut 2 project. I’ll have the basics in place to move on to another application,” he said.