A research team at the University of Toronto is contributing to stellar views of the cosmos by employing innovative telescope technology.

A new from U of T’s highlights the work of a trio of researchers who built, tested and installed brand new equipment that can enhance precision and improve resolution by correcting for distortions in the earth’s atmosphere.

PhD candidate Robin Swanson is part of a research team at U of T’s Dunlap Institute for Astronomy and Astrophysics that has developed telescope technology to enhance precision and improve resolution at large ground-based observatories. (Photo: Dunlap Institute)

The team includes, a PhD candidate in the ������Ƶ of Computer Science and the Dunlap Institute, whose research focuses on combining astronomical instrumentation with machine learning and computer vision tools.

He worked alongside two Dunlap Institute colleagues: PhD student Jacob Taylor and Associate Professor Suresh Sivanandam of the David A. Dunlap ������Ƶ of Astronomy and Astrophysics.

The technology they used, known as adaptive optics, draws on adjustable mirrors on telescopes and specialized cameras to measure and compensate for atmospheric turbulence, which the researchers say is key in unlocking the full potential of ground-based telescopes in large observatories.

Together they built a near-infrared camera and developed machine learning algorithms to improve sky coverage and robustness of adaptive optics systems.

With these tools the team expects “up to ten times improvement” in sky coverage over traditional systems, according to Sivanandam.

Using predictive control, Swanson says the team can estimate what the atmosphere will look like in the future, by using historical conditions that are already present in adaptive optics (AO) systems and newer methods from computer vision and machine learning.

“Because all AO systems have a small amount of latency, there’s a difference between the atmosphere that we measure and the one that we correct for. So, if we can predict what the atmosphere will look like in the future, we can apply that correction instead and get better performance out of our AO system,” Swanson explains.

“Not only can our method be applied to new telescopes that are being developed now, but ideally they could also be used in previous telescopes and upgrade their performance without any costly hardware changes,” he adds.

Alongside colleagues at the University of Arizona and Arizona State University, the U of T team installed the system last fall at the MMT Observatory in Tucson, Arizona and the telescope will be available for scientific observations later this year.

Check out their full story and go behind the scenes in the short documentary below.