In a major leap for space technology, researchers at Julius-Maximilians-Universität Würzburg (JMU) have successfully demonstrated the first-ever use of Artificial Intelligence (AI) to directly manage a satellite’s attitude, its orientation in space, during orbit. The milestone was achieved through the university’s 3U nanosatellite, InnoCube, which completed an AI-controlled maneuver between 11:40 and 11:49 a.m. CET on October 30, 2025.
Unlike conventional satellite controllers that rely on rigid, pre-programmed algorithms, the Würzburg team developed an AI controller based on Deep Reinforcement Learning (DRL), a form of machine learning that enables the system to learn optimal strategies through trial and feedback in a simulated environment. Before launch, the AI was trained extensively in a highly realistic simulation and later uploaded to InnoCube’s flight system.
A new Artificial Intelligence (AI)-based altitude controller for orbiting satellites has been successfully tested by researchers at the University of Würzburg, Germany, marking the first time in history AI has been used for the purpose.
The team utilized a branch of machine… pic.twitter.com/cZ7eEr0QY1— CGTN America (@cgtnamerica) November 11, 2025
The experiment was a resounding success: the AI autonomously adjusted the satellite’s orientation from its initial position to a target attitude, proving its ability to make real-time decisions in space. Researchers highlighted that overcoming the “Sim2Real gap” ensuring that models trained in simulation perform reliably in actual orbit, was among their greatest challenges.
You May Like To Read: Iraq Records Over 55% Turnout in Parliamentary Elections
This breakthrough paves the way for fully autonomous spacecraft operations, potentially revolutionizing future deep-space exploration, reducing mission costs, and accelerating the development of more adaptive, resilient satellite control systems.
Check out our latest video:
