Could next-gen artificial intelligence help map out future space outposts?

Could next-gen artificial intelligence help map out future space outposts? Robert Egan Senior Editor Planning a successful trip can be challenging, especially when your destination is in outer space. Even before liftoff, space missions require extraordinary amounts of scientific preparation and data analysis to succeed, which can take years to complete. However, as space agencies race to establish a permanent presence on the moon, using artificial intelligence to combine such information into a single system could make mission planning much more efficient. One such system, developed by Gokce Inal, a Ph.D. candidate in civil, environmental and geodetic engineering at The Ohio State University, is an AI framework designed to help scientists evaluate potential lunar landing sites, identify dangerous hazards future crews might face, and scour new regions for scientifically valuable geologic features. "AI systems are going to become critical tools for space exploration that will help humans analyze the millions of images and data sets that future missions will require," she said. "This work is meant to serve as a bridge between those two worlds, AI and planetary science." Landing on the moon is more difficult today than it ever was, but preliminary results from Inal's model, dubbed "LUMINA," show a strong ability to make complex lunar observations as well as generate scientifically meaningful mission commentary for candidate landing sites. Such capabilities could help mission planners locate areas abundant with water ice and other resources. An active assistant for mission planners Built on a previously developed lunar vision-language foundation model called LLaVA-LE that could describe lunar terrain and answer scientific questions, LUMINA extends the team's research by using a framework that can plan and carry out multistep analyses, evaluate potential lunar landing sites and explain the evidence behind its recommendations, said Inal. While she created LUMINA to advance lunar mission objectives, Inal envisions the AI model not as a semiautonomous machine, but as a decision-support system designed to work in tandem with mission planners to make final judgments. "Most AI tools are passive tools," she said. "You ask a question and get an answer, but we're moving toward an active scientific assistant that can work through a problem alongside the experts." Inal presented her findings this month at The 58th Annual Meeting of the Division for Planetary Sciences (DPS 58) of the American Astronomical Society. Mapping worlds beyond the moon One of this work's main goals is to establish a pathway for agentic systems to be used in support of future space exploration, but LUMINA's capabilities extend beyond only data interpretation for the lunar surface. Instead, the model could have research and exploration applications for mapping the surface of many other planetary bodies, including Earth, said Alper Yilmaz, co-author of the research and a professor in civil, environmental and geodetic engineering at Ohio State. "There aren't many GPT-like map models being developed for space exploration, so our effort is aimed at making one generalizable to other surfaces, like Mars or some asteroids," he said. As new data sets become more advanced, more efficient tools like this model will be needed to help mission planners search them—otherwise trying to discern the best places for future crews to land or investigate would be like looking for a needle in a haystack, said Yilmaz. "People are going to be landing in an extremely harsh environment, so it's worth having another set of eyes that can verify and question things and help us communicate better," he said. As the system continues to develop, the team plans to test it on new high-quality lunar data and further evaluate how well it supports mission planning. "NASA's Artemis program creates important opportunities for our research," she said. "As new lunar data becomes available, their results can improve systems like ours and better support future mission planning." Provided by The Ohio State University