Researchers develop crater-based AI system for lunar navigation

Researchers in Australia have developed a crater-based artificial intelligence (AI) navigation system to help spacecraft determine their position during long-duration lunar mapping missions, El.kz cites Xinhua.

Researchers from Adelaide University's Australian Institute for Machine Learning (AIML) and the Andy Thomas Center for Space Resources developed STELLA (Spacecraft craTer-basEd Localization for Lunar mApping), bringing landmark-based navigation to the Moon, according to a university statement released Wednesday.

"Our system helps a spacecraft detect the craters it sees through its camera, match them with the craters on the lunar map, and use that information to determine its position," said Chee-Kheng Chng from AIML, co-lead author of the study published in the journal Astrodynamics.

Researchers tested the system on a comprehensive and realistic synthetic dataset of more than 15,000 images emulating a year-long lunar orbit under a range of conditions, including challenging lighting and different viewing angles.

STELLA consistently achieved meter-level position accuracy and sub-degree altitude accuracy in most scenarios, the researchers said, adding the system could provide a baseline for further development of vision-based navigation solutions for long-duration orbital missions.

"Because it's not possible to put people in space easily or cheaply, there needs to be a way to find information before you go, and that's where AI comes in," said study co-lead author Sofia McLeod from AIML.

With global interest in lunar exploration growing, accurate location data will be increasingly important, particularly for missions searching for resources such as water on the Moon, McLeod said.