Stanford Engineers Teach Spacecraft to "Dream" Their Way to the Space Station

Stanford researchers have developed an AI system called the Out-of-this-World-Model (OWM) that enables spacecraft to "dream" their way to the space station. The OWM uses a world model, which is an AI architecture that learns the fundamental physics of the environment it's interacting with directly from experience, rather than from equations programmed into it. The system uses reinforcement learning algorithms and simulates potential futures to guide spacecraft during docking maneuvers. The OWM model required only 500,000 iterations to master docking maneuvers, whereas a comparable RL system required 25,000,000 permutations. The model also performed better when docking at a never-before-seen docking port on the ISS and was better able to deal with unexpected situations.

The OWM model uses a library called AstroJAX, which is designed to run on Graphics Processing Units (GPUs), similar to what early AI models were trained to run on, but originally intended for use in video gaming computers. The researchers found that the OWM model could successfully dock with the ISS docking ports about 53% of the time, compared to the RL algorithm's 29% success rate. However, close-up operations appeared difficult for the OWM model - likely because of high penalization weighting of collisions, which the authors state can be adjusted in future iterations.

The OWM model's success in docking maneuvers is a significant step forward in the development of autonomous spacecraft systems. The ability to simulate potential futures and make decisions based on probability rather than fixed rules could lead to more reliable and efficient space operations. However, the model still has limitations, such as its difficulty in handling close-up operations and the need for further refinement. As the number of satellites that need servicing or de-orbiting increases, and as more people are expected to be in space, the development of such AI systems will become increasingly important for the future of space exploration.

AI-generated summary of the original article — see the source for the full story.