In a world-first achievement, researchers at Julius-Maximilians-Universität Würzburg (JMU) have successfully demonstrated an AI-based satellite attitude controller operating autonomously in orbit. The test was conducted on the 3U nanosatellite InnoCube, marking a major step toward fully autonomous space systems.
On 30 October 2025, during a nine-minute satellite pass between 11:40 and 11:49 a.m. CET, the AI controller, trained using Deep Reinforcement Learning (DRL), guided the satellite from its initial orientation to a precise target attitude using reaction wheels. Subsequent trials confirmed the AI’s reliability in controlling the satellite entirely without human input.
The project, called LeLaR (In-Orbit Demonstrator for Learning Attitude Control), aims to develop next-generation autonomous satellite systems. Traditional attitude controllers require months of manual tuning to stabilize satellites and point them toward cameras, sensors, or antennas. The Würzburg approach uses AI to learn the optimal control strategy in simulation before deployment, dramatically speeding development and offering adaptive capabilities in orbit.
“We have achieved the world’s first practical proof that a satellite attitude controller trained with DRL can operate successfully in orbit,” said Dr. Kirill Djebko, part of the LeLaR team.
A central challenge was bridging the Sim2Real gap—ensuring that a controller trained in a simulated environment performs accurately in the unpredictable conditions of real orbit. The successful test proves that AI can reliably handle this critical task.
Also Read: Microbe That Could Turn Martian Dust into Oxygen
According to the team, the achievement opens doors for future autonomous missions, including interplanetary or deep-space operations where communication delays make human control impractical. “This is a major step toward full autonomy in space,” said Professor Sergio Montenegro.
Funded by the German Federal Ministry for Economic Affairs and Energy and managed by the German Space Agency (DLR), the LeLaR project represents a milestone in AI-driven space technology. As autonomous systems become more sophisticated, researchers hope these advances will lead to smarter, faster, and more cost-effective satellite operations in orbit—and beyond.
Story Source: Julius-Maximilians-Universität Würzburg