NASA could soon use a solution powered by artificial intelligence for a faulty instrument in its Solar Dynamics Observatory, which observes solar activity in extraordinary detail.
Why it matters: The powerful solar storms (bursts of solar plasma and charged particles) can damage satellites in orbit and even cause major problems for energy networks on Earth. And SDO, which can detect solar storms in real time, is a key part of that.
What did you do: The SDO instrument in question is known as MEGS-A, and was designed to monitor levels of ultraviolet radiation, which correlate with a balloon in Earth's outer atmosphere that can damage satellites in near-Earth orbit.
A deep learning network that IBM created in 2018 can soon replace the failed instrument by inferring what levels of ultraviolet radiation that instrument would detect based on what the other SDO instruments are observing at a given time, says Axios the AI consultant. from NASA Graham Mackintosh.
"Imagine you have an orchestra playing, and at some point, for some reason, halfway, one of the musicians who played the violin stopped and left, could you know enough about the way he played the music? open and recreate what you think that musician would have played if that musician was still sitting in the chair, that's something we did. "
– Graham Mackintosh, IA consultant at NASA
Although NASA is not yet using the solution operationally, the results are promising, Mackintosh added.
What to see: AI models like this one could also be used for other future missions, said Mackintosh. Instead of loading three instruments on a satellite to measure different aspects of the space environment, you could launch two and use the collected data to infer the information that would have been measured in a third party.