Hunt for life in space on demands too much bandwidth

Imagine you’ve got a revolutionary new space telescope which can capture more precise data for your project to find life in space around our nearest neighbour, Alpha Centauri, but the files are too big to download … What do you do?

In the case of Sydney University’s shoestring-budget TOLIMAN orbital telescope, it’s yet another cause for innovation.

“So our telescope only passes over a ground station a few times each day, and we have to bucket down as many bits as we possibly can,” says Professor Peter Tuthill from the Institute for Astronomy and School of Physics. “That process depends on money. It depends on how much you spend on your radio and powering it.”

It’s not just a problem in the search for habitable planets orbiting the three-star system that is Alpha Centauri.

It’s also a severe chokepoint for the burgeoning low-Earth observation industry monitoring the ecological impact of climate change, providing timely fire and flood disaster data, policing ocean fishing grounds and measuring the weather in ever greater detail. 

Sydney-based computing company Spiral Blue thinks AI-enhanced accelerated processing is the solution. And the SmartSat Cooperative Research Centre has awarded the TOLIMAN project partnership a grant to help prove the technology’s effectiveness.

Cosmos: TOLIMAN looks for life in space

“This is a very active field at the moment,” says Tuthill. “People are building more and more capable radios. But once you hit the limit with how much you can possibly get down, well, the next challenge is sifting the data so you only get what you need.”

The TOLIMAN telescope images a unique flower-shaped “fingerprint” of its target stars 10 times every second. Minute changes in this “fingerprint’s” composition are expected to point to the presence of any rocky planets within their habitable “Goldilocks” zones. 

Returning every pixel of every image promptly would require a far larger and more expensive radio than the project can afford. So the alternative is to put more processing power in space with the telescope itself.

“Spiral Blue’s technologies allow us to circumvent some of those decisions we would otherwise have to make,” Tuthill adds. “Now those decisions can be made onboard – such as selecting what features of the data should be downloaded, or sending us already refined and processed results.”

But integrating such technology (dubbed edge computing) with a satellite requires a delicate juggling act of its own.

“We need to fly high-performance computing elements able to withstand the space environment and operate efficiently on the kinds of tight power budgets up there,” Tuthill explains. “This collaboration with Spiral Blue allows us to stream data far better than we were expecting to do just a couple of years ago.”