IBM, NASA Launch AI Model to Help Map Ice, Craters on Moon

Scientists may soon be analyzing the Moon's surface and search for water ice, maps of its craters, and even the safest spots to land future missions more quickly, thanks to a new open-source AI model by IBM and NASA.

The NASA-IBM Lunar Foundation Model is trained on over 30 layers of data collected by nine different instruments across four NASA missions, including the Lunar Reconnaissance Orbiter. This lunar model is part of a broader series of open foundation models called Prithvi, developed by IBM and NASA, used in fields ranging from Geospatial analysis to weather and environmental research.

This new lunar model will enable the identification of essential lunar surface features that researchers have historically relied on to discover and map these using many maps and images compiled together. It can pinpoint possible ice deposits, estimate where possible water ice may reside and map features such as craters or volcanic terrain over vast expanses of lunar terrain.

Performance tests demonstrate that this model has discovered key lunar features up to 23 percent more effectively than conventional methods already in wide use, say IBM and NASA. This advancement may be significant as space agencies prepare for increasingly complex missions and struggle with processing greater amounts of information generated from these observational datasets.

Water ice deposits represent some of the most desirable objectives, as it may potentially offer a resource for future astronauts. It may also be split into hydrogen and oxygen as elements used for future propulsion systems and to maintain vital functions on future missions, minimizing the volume of material sent from earth to carry upon missions.

The map produced is also an aid to safer missions; mapping volcanic features, craters and terrain helps scientists analyze possible landing positions and pinpoint hazardous regions for both spacecraft and humans traveling there.

These developments coincide as NASA steps up its program on the Moon, Artemis, and its ambitious plans for sustained humans on the moon's surface. NASA intends to return humans to its surface in 2028, following tests involving its use as part of missions with a view to enabling similar human exploration of mars.

The tool is open source, so it enables scientists wider access to a program developed and used for processing lunar data. The work from NASA-IBM partnership is an increasing reliance upon AI for the purposes of exploration and use on space missions; machine learning systems can aid with turning large amounts of scientific data, including a multitude of map-like outputs, into understandable information to be displayed more rapidly than conventional techniques.