Planning a space expedition is an enormously complex task. Even before launch, vast amounts of scientific data must be collected and analyzed, and this can take years. As space agencies seek to establish a presence on the Moon, combining such data into a single system with the help of artificial intelligence (AI) could significantly speed up mission preparation.
One such tool was developed by Gokce Inal, a graduate student in the Department of Civil, Environmental and Geodetic Engineering at Ohio State University. Her AI-based system is designed to assess potential landing sites on the Moon, identify hazards that future crews may encounter, and find new areas with geological features of scientific value.
"AI systems will become critically important tools for space exploration; they will help people analyze the millions of images and datasets that future missions will require," the researcher says. "This work is intended to become a bridge between two worlds - AI and planetary science."
The model is called LUMINA. Preliminary results show that it is capable of performing complex observations of the Moon and providing substantive commentary on possible landing sites. This could help mission planners find areas rich in water ice and other resources.
LUMINA is built on the earlier LLaVA-LE model, which could describe lunar terrain and answer scientific questions. The new system can plan and carry out multi-step analysis, evaluate landing sites and explain what its recommendations are based on.
Inal stresses that she does not regard the model as a semi-autonomous machine. In her conception, it is a decision-support system that works together with mission planners, while the final conclusions remain with humans. "Most AI tools are passive: you ask a question and get an answer. We are moving toward an active scientific assistant that works through a problem together with experts," she explains.
The results were presented at the 58th annual meeting of the Division for Planetary Sciences of the American Astronomical Society (DPS 58).
LUMINA's capabilities are not limited to the Moon. According to co-author Alper Yilmaz, a professor at the same university, the model could be useful for mapping the surface of other planets, including Earth, Mars and some asteroids. "Few map models like GPT are being created for space research, so our goal is to make a universal one," he notes.
As data volumes grow, more efficient tools will be needed to search them. Otherwise, Yilmaz believes, finding the best places for landing or research will be like looking for a needle in a haystack. "People will be landing in an extremely harsh environment, so it is worth having another pair of eyes that can check and challenge decisions and help explain everything better," he says.
The team plans to test the system on new high-quality lunar data and further assess how well it supports mission planning. "The Artemis program creates important opportunities for our research. As new lunar data become available, their results can improve systems like this and better support the preparation of future missions," Inal concludes.