Science

AI can schedule telescope observations like an expert astronomer

By cpstestersofficial@gmail.com · August 6, 2026 · 5 min read

Every night, astronomers face the same puzzle, and this year an AI system stepped in to help solve it.

Weather shifts, the moon brightens or fades, and the air itself blurs starlight in ways that change hour to hour.

Pick the wrong target, and a telescope wastes a night nobody gets back. The AI system matched the astronomers who usually make that call.

Alex Drlica-Wagner, an astronomer at Fermilab and a professor at the University of Chicago, led the project with Aravindan Vijayaraghavan, a computer scientist at Northwestern University.

Both work through the National Science Foundation-Simons Foundation AI Institute for the Sky, a research group nicknamed SkAI.

Every clear night is a gamble

Picking a target isn’t only about finding something interesting to look at. It’s about spending scarce time well.

Astronomers sometimes wait months for a turn on a major telescope.

A badly aimed shot can come back blurry or washed out by moonlight, hiding the faint or distant objects it was meant to catch. Redoing that shot might not be possible again for months.

“Large telescopes are national or international resources,” Drlica-Wagner said.

“Many astronomers around the world want time to use these telescopes, and that time is limited. If everyone could use their time more efficiently, then the community will be able to do more science.”

The computer learned from old decisions

Drlica-Wagner studies large sky surveys. Vijayaraghavan builds machine-learning systems.

Together with their teams at SkAI, they built a system that learns scheduling on its own, rather than following rules astronomers wrote down by hand over the decades.

The experts trained the sysytem on years of past observations from the Dark Energy Survey, a project that scans the sky with a large camera bolted to a telescope in Chile.

“We trained the model on years of historical observations by showing it where the telescope was pointing at one moment and asking it to predict the next observation,” Drlica-Wagner said.

“Then we compared its prediction to what astronomers actually did and asked it to correct its mistakes.”

After enough rounds of that correction, the system started picking targets on its own.

It weighed moonlight and shifting atmospheric conditions the way a human scheduler would. No one had to write those rules down for it.

Real nights in the Chilean mountains

This past spring and summer, the system ran two real observing sessions on the 13-foot (4-meter) Víctor M. Blanco Telescope at Cerro Tololo Inter-American Observatory in Chile.

It scheduled time on the 570-megapixel Dark Energy Camera, built by the Department of Energy and mounted on the Blanco.

The system didn’t just hand over one fixed plan for the night. It adjusted its choices as clouds rolled in or the moon rose, the same way a human scheduler would mid-shift.

Paul Chichura, a SkAI postdoctoral researcher, along with University of Chicago doctoral student Rachel Hur and NOIRLab scientist Guillermo Damke, ran the deployment on-site.

“This is an important milestone toward more autonomous observatories,” Drlica-Wagner said.

“One of the main achievements is that we set up all the infrastructure needed to deploy this self-driving telescope on a national observatory.”

“Currently, I would say its performance is comparable to a human’s ability. As the next step, we plan to teach the computer to do a better job than a human.”

What a busier sky will need

More telescopes are coming online, including the NSF-DOE Vera C. Rubin Observatory. Together they will gather far more data each night than any one scheduler, human or computer, has had to handle before.

A faster, adaptive scheduling system could help several telescopes work together and get more out of every hour of dark sky, a goal the SkAI team is already pursuing.

“It is exciting to see ideas from AI and reinforcement learning brought to telescope scheduling, where every decision must balance changing conditions and scarce observing time,” Vijayaraghavan said.

“Developing intelligent scheduling systems for astronomical surveys also raises fascinating new machine learning problems, and we are excited to continue exploring them through this project.”

Nobody yet knows whether a computer can out-schedule a trained astronomer, only that it can keep up with one. The SkAI team’s next step is to test strategies a human scheduler might never think to try.

“If we can automate this technical operational task so it requires less human effort, then astronomers can have more time to think about more scientifically interesting problems and focus on discovery,” said Drlica-Wagner.