Teaching Space Robots to Think: Naval Research Lab’s Reinforcement Learning Breakthrough

This edition of Fed Gov Today with Francis Rose, takes you to the cutting edge of robotics in space. Dr. Samantha Chapin, Space Roboticist, and Dr. Kenneth Stewart, Computer Research Scientist at the Naval Research Laboratory (NRL), share how their team is using reinforcement learning — a flavor of machine learning inspired by dog training — to teach robots to operate autonomously in space. They explain why human time in space is so precious, how reinforcement learning allows robots to adapt in unpredictable environments, and why the International Space Station is the perfect proving ground for these complex experiments.Francis and his guests explore the significance of training algorithms in simulation, using open-source NASA tools, and then successfully testing them aboard the ISS on Astrobee robots — a first step toward enabling autonomous space assembly, servicing satellites, and performing complex repairs. The conversation dives into future goals, from CubeSat experiments to advancing autonomous robotics that can operate in space without constant human oversight, helping build and maintain the next generation of massive space structures.Subscribe and listen to the Fed Gov Today Podcast anytime on Apple Podcasts, Spotify, or at FedGovToday.com.