AI pilots a free-flying robot aboard the ISS for the 1st time (scienceclock.com)

In a groundbreaking experiment, researchers from Stanford University have successfully demonstrated AI-assisted autonomous navigation for the Astrobee robot aboard the International Space Station (ISS). This marks the first time that artificial intelligence has been utilized to control a robot in the ISS environment, which presents unique navigation challenges due to its confined spaces and reduced computing power compared to terrestrial settings. The team developed a machine-learning model that helps the Astrobee robot generate a preliminary flight plan, called a "warm start," based on previously solved paths, allowing for 50-60% faster trajectory planning while ensuring safety protocols are maintained. This advancement is significant for the AI and space robotics community, as it paves the way for increased autonomy in future space missions. With growing distances from Earth to targets like Mars, real-time human control of robots will become impractical due to communication delays. By enabling robots to navigate independently within strict safety parameters, this AI-enhanced approach not only improves operational efficiency aboard the ISS but also lays the groundwork for more complex, long-duration missions where continuous oversight may be impossible. The successful demonstration indicates that AI can be an essential tool in addressing the unique constraints of space environments, marking a step forward in the exploration of autonomous robotics in space.