Advertisement
Australia markets open in 5 hours 34 minutes
  • ALL ORDS

    8,262.40
    +56.30 (+0.69%)
     
  • AUD/USD

    0.6761
    -0.0025 (-0.37%)
     
  • ASX 200

    8,017.60
    +58.30 (+0.73%)
     
  • OIL

    81.86
    -0.35 (-0.43%)
     
  • GOLD

    2,428.20
    +7.50 (+0.31%)
     
  • Bitcoin AUD

    93,901.87
    +5,193.31 (+5.85%)
     
  • CMC Crypto 200

    1,318.07
    +49.12 (+3.87%)
     

DARPA and Slingshot build system to detect ‘wolf in sheep’s clothing’ adversary satellites

Image Credits: Andriy Onufriyenko / Getty Images

The number of satellites on low Earth orbit is poised to explode over the coming years as more mega-constellations come online. This will create new opportunities for bad actors to hide weapons or spy satellites among their fleet.

That’s according to the Defense Advanced Research Projects Agency (DARPA), which tasked space startup Slingshot Aerospace with developing a new tool that could pick out potentially nefarious satellites purposefully hiding in mega-constellations. That system is called Agatha — named after one of the precogs in the film "Minority Report" — and it could provide a huge boon for national space security.

DARPA officially selected Slingshot for the program — called Predictive Reporting and Enhanced Constellation Objective Guide (that’s right, PRECOG) — last March, and the work formally concluded this January. The company was awarded around $1 million for the work, according to government contracting database HigherGov.

Slingshot researchers generated 60 years’ worth of synthetic constellation data with which to train Agatha, so that the system could detect minute differences in satellite behavior, using those differences to deduce the satellite’s true operational directives.

ADVERTISEMENT

As it turns out, there are “lots of little tells that add up to something much bigger,” Slingshot’s director of data science, Dylan Kesler, said in a recent interview. That could be small changes in the mass of the satellite, which affects its station keeping, or if the satellite communicated differently with Earth, or if it was always oriented in the same direction (versus all the other satellites in the constellation).

While Agatha was trained on simulations, Slingshot eventually tested it against real-world constellations by identifying non-nefarious outlier satellites in operators’ existing fleets. The program, which is running on Slingshot’s space domain awareness platform today, now culls data from the company’s own Global Sensor Network and its Seradata database, as well as other public and proprietary sources.

Other startups have generated buzz for their plans to develop rendezvous tech to gather intelligence on adversary satellites. Slingshot’s Agatha solves the crucial question of how these operators will identify those satellites in the first place. This question becomes all the more imperative as countries like China announce plans to launch multiple mega-constellations over the coming decade.

“Historically, when there weren't these 10,000 or 15,000 mega-constellations, it’s reasonable that a human or a team of humans could look at orbital trajectory data and other sources of data and make an informed assessment of which satellites to go after,” Slingshot’s VP of strategy and policy, Audrey Schaffer, said. “But as the growth and activity in space just increases exponentially, it's going to become impossible for a human to really sift through all of this data without the help of tools like Agatha.”