Skip to main content
European Commission logo
Research and Innovation

AI detective discovers unknown stars and probes fundamental laws of the universe

Bursts of radio waves coming from outside our solar system have puzzled scientists since their discovery in 2007. By combining a radio telescope with machine learning software, the EU-funded MeerTRAP project was able to pinpoint the origin of these signals. As well as uncovering new celestial bodies, the system is helping physicists test their theories.

©vchalup #694745 source: 2023

PDF Basket

No article selected

Earth is bombarded by radio signals from space every day. The source of most of them is unknown. While most people won’t take much notice, for astronomers, these signals offer a way to study distant objects and deduce what object or event might have sent them our way.

In 2007, the Parkes radio telescope in Murriyang, Australia, detected a short burst of radio emission that seemed to come from a source well outside our galaxy – something that had never been seen before.

It wasn’t until 2013 that similar signals were detected, and scientists coined the name ‘fast radio bursts’ (FRBs). “We were now confident that they were real,” says Ben Stappers, MeerTRAP project coordinator and head of the Pulsars, Exoplanets and Transients Group at the University of Manchester in the United Kingdom.

FRBs are intense, millisecond bursts of radiation coming from distant galaxies. It’s thought that over 10 000 FRBs reach Earth every day, but their origin is still shrouded in mystery. A popular theory is that they are emitted by neutron stars, elusive remnants of dying stars with exceptionally strong magnetic fields.

The EU-funded MeerTRAP project used a super-sensitive radio telescope in South Africa known as MeerKAT, coupled with state-of-the-art hardware and software to search the skies for FRBs and try to understand more about these extraterrestrial signals.

Eyes to the skies

Because these signals can come from so far away in space, astronomers can find it difficult to tell exactly which galaxy they came from.

The two main advantages of the MeerKAT telescope are its high sensitivity and its ability to localise sources of radiation. This allowed the team to find more FRBs and start building up an accurate map of their origins.

“We predicted at the time that we would find something like one fast radio burst every two weeks we were looking at the sky,” adds Stappers. “That prediction seems to be holding well, as we have discovered around 45.”

Using AI to hunt for clues

The data produced by these telescopes is so extraordinarily large that long-term storage is either expensive or impossible. The MeerTRAP team needed to quickly sift through the data to uncover potential leads.

To do this, the researchers developed sophisticated software, including a machine learning tool that quickly scans the data and decides whether a sample may contain the signature of an FRB or not.

The team also developed a tool that allows them to make an image of the sky using a small time slice of data when the burst occurred. This gives a precise location of the FRB. They are now working to use these signals as cosmological probes: a way to test theories in physics such as gravity, and to understand mysterious objects in space.

Pulsars fast and slow

The team also detected FRBs within our own galaxy, the Milky Way. These signals most likely came from pulsars: fast-rotating and highly magnetised neutron stars that fire out radiation intermittently into space. So far, the team have found more than 85 such objects.

The project also advanced our understanding of these cosmic entities. Before the project started, pulsars were thought to spin with a period of one and a half milliseconds up to around eight seconds. Since then, this figure has grown, but the MeerTRAP project discovered an object rotating with a very slow spin period of 76 seconds, an unexpected result that was published in the journal ‘Nature Astronomy’.

“A lot of people are surprised that you can have an object that spins that slowly and still be emitting radio waves,” remarks Stappers.

The project opened up many more questions to be answered. The team are looking forward to taking lessons learned in this project and applying them to the Square Kilometre Array, a powerful new pair of telescopes being built in Australia and South Africa.

”Thanks go to the European Research Council, SARAO and MPIfR, without whom this exciting project would not have been even remotely possible,” says Stappers. “I’d like to pay tribute to the talented group of early career researchers who made MeerTRAP a fun and successful project,” he adds.

PDF Basket

No article selected

Project details

Project acronym
Project number
Project coordinator: United Kingdom
Project participants:
United Kingdom
Total cost
€ 3 488 956
EU Contribution
€ 3 488 956
Project duration

See also

More information about project MeerTRAP

All success stories