This article is more than 1 year old
Starlink decoded for use as GPS alternative – without Elon Musk's help
Told no by SpaceX, uni eggheads went ahead by themselves
Researchers rebuffed by SpaceX have taken matters into their own hands and reverse-engineered Starlink's satellite signal for potential use as a GPS alternative.
University of Texas Austin professor Todd Humphreys and his team claim in a non-peer reviewed paper that they've managed to decode Starlink downlink signals in the 10.7 to 12.7 GHz band detailed enough that they were able to locate a Starlink receiver to within 30 metres.
That's not as accurate as traditional GPS, but it's noteworthy since Humphreys' team managed to do it without any SpaceX help. If Musk and company decided to play ball, Humphreys told us, Starlink positioning could become more accurate than GPS with little work.
Humphreys's team has been working on the technique since 2020, when talks between the US Army, his team at UT and and SpaceX ended with a decree from His Muskiness: "every other [low Earth orbit] communications network has gone into bankruptcy, and so we [SpaceX] have to focus completely on staying out of bankruptcy. We cannot afford any distractions," Humphreys said Musk had decided.
The US Army was interested in using Tesla's LEO satellite constellation as an alternative to GPS signals because of their aforementioned improved accuracy and resistance to jamming. But SpaceX guarded its signal design closely, revealing absolutely nothing about how it worked during the 2020 talks.
When Musk slammed the door on UT and the Army, Humphrey's team bought a receiver, began streaming videos from it and built "basically a little radio telescope to eavesdrop on their signals," he said.
Humphreys' team wasn't interested in the customer data coming from Starlink satellites, and instead focused on synchronization services that help satellites and receivers communicate. If decoded, such signals could be used to get pseudorange measurements used for positioning.
Using an approach called blind identification, the researchers found four sync sequences that could be exploited to use as a positioning service. They were also able to determine that Starlink satellites use orthogonal frequency-division multiplexing (OFDM) to encode transmissions, an apparent first.
- Biden wants SpaceX to beam internet to Iran amid uprising
- Starlink, shot by both sides in Ukrainian fracas, lives to fight on
- US to relax restrictions for tech companies in Iran
- Satellite operators want option to exceed deorbiting rules
"Insofar as we are aware, blind identification of operational OFDM signals, including exact determination of synchronization sequences, has not been achieved previously," the researchers said in the paper [PDF].
According to Humphreys and his team, it would take little more than a software update to make Starlink satellites capable of serving as a GPS alternative, or even successor. "It could be done immediately, with no modifications to the satellites," Humphreys told The Register.
Unfortunately, it'd still be limited to the 30 meters the researchers have already demonstrated.
But reducing that number wouldn't be too difficult, Humphreys told us. He said SpaceX could publish a clock correction for each satellite - something that is currently done for GPS - which would reduce errors to under one meter.
"This wouldn't take [SpaceX] much work, and would require only modest changes to the satellite software," Humphreys told us. Whether that software update is even a twinkle in Elon's eye is unknown - we've reached out to SpaceX to learn more about its plans but haven't heard back. ®