Scientists from Northwestern University, the University of Chicago, and Fermilab have created a novel artificial intelligence (AI) tool that autonomously decides the optimal positioning of a telescope.
Each night, astronomers carefully assess changing weather, moonlight, and atmospheric conditions before deciding where to point a telescope. This constant balancing act helps maximize the scientific value of every precious hour under dark skies.
Developed at the National Science Foundation (NSF)-Simons Foundation AI Institute for the Sky (SkAI, pronounced “sky”), the researchers successfully employed the AI system to organize observations with the 570-megapixel Dark Energy Camera (DECam), constructed by the U.S. Department of Energy (DOE), which is mounted on the NSF Víctor M. Blanco 4-meter Telescope located at the Cerro Tololo Inter-American Observatory (CTIO) in Chile.
Keep a PDF of this page
The system not only produced an observing schedule but also modified that schedule in real time as environmental conditions evolved. By automating routine scheduling tasks, this advancement will enhance the ability of telescopes to gather the most accurate data possible.
This is an important milestone toward more autonomous observatories. 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.
Alex Drlica-Wagner, Project Lead, University of Chicago
Drlica-Wagner is a researcher at Fermilab and serves as a professor of astronomy and astrophysics at the University of Chicago. Wagner collaborated on the project with Aravindan Vijayaraghavan, who is an associate professor of computer science at Northwestern University’s McCormick School of Engineering.
The on-sky deployment at the Blanco telescope was carried out by Paul Chichura, a SkAI postdoctoral associate; Rachel Hur, a Ph.D. student at the University of Chicago; and Guillermo Damke, an associate scientist at NSF’s NOIRLab. Drlica-Wagner, Vijayaraghavan, Chichura, and Hur are all integral members of SkAI.
Determining the direction in which to aim a telescope involves more than simply locating an intriguing object to study. It also requires optimizing every precious minute of observing time. Astronomers may wait for months to gain access to a significant telescope. If a telescope is not properly aligned, it may produce images that are less clear or overly illuminated by moonlight, complicating the detection of faint or distant objects. The chance to repeat a failed observation could be delayed for several months.
Large telescopes are national or international resources. 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.
Alex Drlica-Wagner, Project Lead, University of Chicago
Drlica-Wagner integrated his knowledge of extensive astronomical surveys with Vijayaraghavan’s proficiency in advanced machine learning to enhance the efficiency of the observation process. Together with the teams at SkAI, they created a deep-learning scheduling system. Instead of coding the AI with rules that astronomers have established over many years, the researchers allowed the system to learn independently.
The scientists trained a deep-learning model using historical data from the DOE-funded Dark Energy Survey, which examines the night sky through a large camera attached to the Blanco Telescope.
“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. Then we compared its prediction to what astronomers actually did and asked it to correct its mistakes. After repeating this process many times, it learned how to schedule observations without being explicitly taught how the brightness of the moon, the atmospheric conditions, or the many other factors affect the quality of astronomical observations,” said Drlica-Wagner.
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. 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.
Aravindan Vijayaraghavan, Associate Professor, Computer Science, Northwestern University
During the previous spring and summer, the intelligent scheduling system successfully completed two observing runs on the Blanco Telescope, which is recognized as one of the most productive astronomical facilities globally.
The primary objective of this initial deployment was to enable the AI to perform comparably to human schedulers. The subsequent aim for the team is to instruct the AI not only to replicate human decision-making but also to enhance it. The AI has the potential to significantly increase the efficiency of telescopes by investigating observing strategies that humans may overlook.
As next-generation telescopes, such as the NSF-DOE Vera C. Rubin Observatory, start generating unprecedented volumes of astronomical data, intelligent scheduling systems could assist companion telescopes in responding more effectively and maximizing the scientific value of each observation run.
“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.
SkAI, spearheaded by Northwestern University, is a National AI Research Institute that receives joint funding from the NSF and the Simons Foundation. This institute unites researchers from astronomy, artificial intelligence, and associated disciplines to create reliable AI tools that enhance scientific discovery, promote state-of-the-art astronomical surveys and instruments, and educate the upcoming generation of interdisciplinary scientists.
The NSF Víctor M. Blanco 4-meter Telescope Observing with an AI Scheduler
An artificial intelligence system has for the first time successfully planned and adapted observations on a national facility, demonstrating a new way to make the most of scarce observing time. Above, the Milky Way arcs over the NSF Víctor M. Blanco 4-meter Telescope at the U.S. National Science Foundation Cerro Tololo Inter-American Observatory in Chile. Video Credit: CTIO/NOIRLab/NSF/AURA/P. Horálek (Institute of Physics in Opava)