AI Tool Deployed to Automate Telescope Observations in Chile

Every night beneath dark skies, astronomers face a relentless balancing act. Before pointing a telescope toward the cosmos, researchers must carefully assess shifting weather patterns, the intensity of moonlight, and fluctuating atmospheric conditions. Squeezing maximum scientific value out of every precious hour requires constant adjustments.

Now, a collaborative team of scientists from Northwestern University, the University of Chicago, and Fermilab has developed a new artificial intelligence tool that automatically determines where a major research instrument should point. The breakthrough brings automated scheduling directly to a national observatory, marking a significant shift in how large-scale astronomical surveys manage their scarce resources.

Deploying Self-Driving Telescope Infrastructure at Cerro Tololo

The newly engineered system was developed within the National Science Foundation-Simons Foundation AI Institute for the Sky, widely known as SkAI. Researchers successfully integrated the technology to organize observations with the 570-megapixel Dark Energy Camera, fabricated by the U.S. Department of Energy. This powerful camera is mounted on the NSF Víctor M. Blanco 4-meter Telescope at the Cerro Tololo Inter-American Observatory in Chile.

Unlike traditional static observation schedules, the intelligent tool generates an initial plan and dynamically adapts it in real time as environmental conditions evolve. During the spring and summer, the system completed two successful observing runs on the Blanco Telescope, which stands as one of the most productive astronomical facilities in the world.

“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

The on-sky deployment at the Blanco telescope was executed by Paul Chichura, a SkAI postdoctoral associate; Rachel Hur, a doctoral student at the University of Chicago; and Guillermo Damke, an associate scientist at NSF’s NOIRLab. The broader research initiative is co-led by Drlica-Wagner—who serves as a scientist at Fermilab and a professor of astronomy and astrophysics at the University of Chicago—alongside Aravindan Vijayaraghavan, an associate professor of computer science at Northwestern University’s McCormick School of Engineering.

Why Efficient Scheduling Matters Under Dark Skies

Choosing a target for a large telescope involves far more than simply locating an intriguing cosmic object. Observing time on premier facilities is fiercely contested. Researchers frequently wait months for a single opportunity to gather data.

Northwestern University Researchers Link Volcanic Eruptions to Ancient Extinction

A poorly positioned telescope can yield blurred images or washed-out frames flooded by moonlight, obscuring faint or distant targets. When an observation fails, the chance to repeat it might not arrive for months. Because major telescopes function as vital national and international resources, maximizing every minute directly impacts the scientific community’s output.

“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

Training Deep Learning Models on Historical Sky Surveys

To build a scheduling system capable of handling these complex variables, the research teams combined expertise in large astronomical surveys with advanced machine learning. Rather than manually programming the artificial intelligence with decades of rigid rules established by astronomers, the developers allowed the system to learn autonomously through reinforcement learning techniques.

The scientists trained their deep-learning model using years of historical observations generated by the Dark Energy Survey, which maps the night sky using the giant 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.”

Alex Drlica-Wagner, Project Lead, University of Chicago

This iterative training process allowed the software to grasp the subtle interplay between changing weather, lunar illumination, and instrument positioning without human intervention.

“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 of Computer Science, Northwestern University

Next Steps for Autonomous Observatory Operations

While the initial spring and summer deployment successfully proved that the AI could match human scheduling competence, the research team is already looking ahead. The primary objective for the next phase of development is to push the software past human limitations.

By exploring novel observation strategies and complex decision pathways that human schedulers might never consider, the engineering team aims to make next-generation telescopes significantly more efficient.