AI Controls Satellite In Orbit For The First Time In Historic Space Milestone

AI conducts a full in-orbit attitude maneuver

During the satellite pass between 11:40 and 11:49 a.m. CET on 30 October 2025, the AI agent developed at JMU performed a complete attitude maneuver in orbit, controlled entirely by artificial intelligence. Using reaction wheels, the AI brought the satellite from its current initial attitude to a specified target attitude. The AI then had several further opportunities to demonstrate its capabilities: in subsequent tests, it successfully and safely controlled the satellite to the desired attitude.

The LeLaR research team, Dr. Kirill Djebko, Tom Baumann, Erik Dilger, Professor Frank Puppe, and Professor Sergio Montenegro, has thus taken a decisive step towards space autonomy.

The LeLaR project: learning attitude control in orbit

The "In-Orbit Demonstrator for Learning Attitude Control" (LeLaR; German: In-Orbit Demonstrator Lernende Lageregelung) (https://www.uni-wuerzburg.de/en/news-and-events/news/detail/news/artificial-intelligence-from-wuerzburg-controls-satellites-in-orbit/) project aims to develop the next generation of autonomous attitude control systems. Its central focus was the design, training, and in-orbit testing of an AI-based attitude controller aboard the InnoCube (https://www.uni-wuerzburg.de/en/news-and-events/news/detail/news/small-satellite-big-potential/) nanosatellite.

Attitude controllers stabilize satellites in orbit and prevent them from tumbling out of control. They are also used to point the spacecraft in a desired direction, for example, to align cameras, sensors, or antennas with a specific target.

Deep Reinforcement Learning powers the AI controller

What makes this work special is that the Würzburg controller was not built using traditional, fixed algorithms. Instead, the researchers applied a Deep Reinforcement Learning (DRL) approach, a branch of machine learning in which a neural network autonomously learns the optimal control strategy in a simulated environment.

Fast and adaptive control through AI

The key advantage of the DRL approach lies in its speed and flexibility compared to classical control development. Traditional attitude controllers often require lengthy manual tuning of parameters by engineers, which can take months or even years.

The DRL method automates this process. Moreover, it offers the potential to create controllers that automatically adapt to differences between expected and actual conditions, eliminating the need for time-consuming manual recalibration.

Bridging the Sim2Real gap in space

Before deployment, the AI controller was trained on Earth in a high-fidelity simulation and then uploaded to the satellite's flight model in orbit. One of the greatest challenges was bridging the so-called Sim2Real gap, ensuring that a controller trained in simulation is also operational on the real satellite in space.

"A truly decisive success," emphasises Kirill Djebko of JMU. "We have achieved the world's first practical proof that a satellite attitude controller trained using Deep Reinforcement Learning can operate successfully in orbit," he adds.

Tom Baumann explains: "This successful test marks a major step forward in the development of future satellite control systems. It shows that AI can not only perform in simulation but also execute precise, autonomous maneuvers under real conditions."

Building trust in AI for space missions

By successfully demonstrating an AI-based controller in orbit, the Würzburg team has shown that artificial intelligence can be reliably applied in safety-critical space missions.

Frank Puppe is convinced: "This will significantly increase the acceptance of AI methods in aeronautics and space research," pointing out the important role of the simulation model.

Growing trust in such technology is a crucial step towards future autonomous missions, for instance, interplanetary or deep-space missions where human intervention is impossible due to vast distances or communication delays. The AI-based approach could therefore become vital for spacecraft survival.

German Space Agency supports AI innovation

With this experiment, the Würzburg team has reached a major goal in the LeLaR project, which has been funded since July 2024 by the Federal Ministry for Economic Affairs and Energy (BMWE) with around €430,000. The project is managed by the German Space Agency at DLR.

"This success motivates us to extend the technology to new scenarios," says Erik Dilger. The test was conducted aboard InnoCube, a satellite developed in cooperation with Technische Universität Berlin (TU Berlin). InnoCube serves as a platform for innovative space technologies, giving researchers the opportunity to test new concepts directly in orbit.

One such innovation is the wireless satellite bus SKITH (Skip The Harness), which replaces conventional cabling with wireless data transmission. This not only saves mass but also reduces potential sources of failure.

Outlook: towards full space autonomy

This successful in-orbit test establishes the University of Würzburg as a pioneer in AI-driven space systems. The demonstrated AI-based controller represents an important building block for future deep-space exploration. The LeLaR project's results may enable faster and more cost-effective development of new, complex AI-based controllers for a wide range of satellite platforms.

"The next goal is to build on this head start," says Kirill Djebko. "It's a major step towards full autonomy in space," adds Sergio Montenegro. "We are at the beginning of a new class of satellite control systems: intelligent, adaptive and self-learning."

Funding and acknowledgements

The projects LeLaR (funding code/FKZ: 50RA2403) and InnoCube (funding code/FKZ: 50RU2000) are funded by the Federal Ministry for Economic Affairs and Energy (BMWE) based on a decision by the German Bundestag.