Hacker News Reader

Comment by swores | Hacker News Reader

Forgive me if I'm thinking like a TV writer since I don't have the scientific knowledge, hence the question: can GPS signals not be spoofed, if you are in possession of the device you want to trick? I.e. Couldn't they have fake GPS signals right next to a Starlink terminal that's not strong enough to have a wide-reaching effect, but enough for the terminal to believe it's somewhere it isn't and report that false location back to the satellites? (And/or finding a way to hack the terminal itself to send false coordinates?)

I guess even if that's possible, it would be feasible for Starlink to triangulate from their satellites to estimate terminal locations and spot which ones are lying?

View on HN

Spoofing a Starlink receiver's GPS is not going to be very helpful with phase-shifted array antennas. Both the satellite and the terminal need to know each other's position.

https://en.wikipedia.org/wiki/Phased_array

Also you must have a fairly good idea just by knowing which satellite is being used.

Starlink satellites are in a very, very low orbit - which means they "see" a fairly small patch of land. So if you're in eastern Ukraine, currently satellites #32581 and #44197 are above you, with #17994 coming soon over the horizon (numbers invented).

Additionally, they have phased antenna arrays able to narrow-beam your data pretty much to you (this is very useful as you can, among other things, recycle the exact same frequency for another user a short distance away).

Spoof GPS all you want, but Starlink knows you're not actually somewhere else, simply because of which satellites are able to talk to you right now.

>(numbers invented)

all numbers are invented ;)

Or discovered

Some are even imaginary

others are real though

That's not natural.

Let's stay rational please

Obviously what Russia will do is send a spacecraft to bump some of the starlink satellites to point upwards towards space, and then launch satellites in higher orbit to bounce the signal to anywhere in Russia, bypassing the geo-fencing /s.

Spoofing is possible, and Russia is known to use it. But it's unlikely they'd use it here. Spoofing is pretty indiscriminate, so nearby devices would also be affected, and as another commenter pointed out the device knowing it's location is important for it to connect to the network.

https://www.nbcnews.com/news/vladimir-putin/russia-spoofing-...

The spoofing could be local to exactly the device you want to spoof. Just put a very weak sender on top of the antenna.

Agree about the other stuff though, and the points made in other comments that Starlink is sort of a GPS itself.

I assume a working gps is necessary for good signal since orientatiin matters in starlink.

I'm going to call on Cunningham's law here... (I'm not sure if I am right or wrong but I think I'm right and hopefully someone can correct me)

I don't think starlink needs GPS as much as the starlink service is a form of GPS. The base station sees several satellites and the satellites know where they are and so the base station can triangulate where you are, and the whole system requires this precision location mechanism to work, so the 2 way communication is a form of Global Positioning without relying on external GPS data.

Perhaps the base station is also able to use GPS but the starlink communication is also able to do exactly what GPS does so "fake GPS" would do nothing to conceal the base station's location and probably make initialization hard or impossible.

Ehm.

You need a precise clock for any form of GNSS[0] to function. AFAIR StarLink satellites doesn't have precise clock on the board. And anyway, modern GNSS chips are.. $1? Cheaper? No way you wouldn't use GNSS already deployed and working (and you are not paying for it) in a commercial product.

[0] GPS is GNSS, it's just in America GPS became the synonym for GNSS

> You need a precise clock for any form of GNSS[0] to function

How precise? I assume that inaccuracy in the clock produces imprecision in location - how much time inaccuracy produces how much location inaccuracy?

> And anyway, modern GNSS chips are.. $1? Cheaper? No way you wouldn't use GNSS already deployed and working (and you are not paying for it) in a commercial product.

Rather than using the Starlink satellites as an alternative to GPS - get a location fix using both Starlink and GPS, and compare them. Even given the former is going to be significantly less accurate - suppose GPS is accurate to within 5 metres and Starlink is accurate to within 5000 metres - well, if the two differ by 50,000 metres, you know something is going wrong, and the terminal might decide to disable itself in such a situation.

> How precise

GNSS satellites orbits are between 20000km and 30000km. 20kkm is 0.06671s of c, 30kkm is 0.10007s, 5km is 0.000016678. So a very precise, that's why they have an atomic clock on board instead of cheap quartz one.

> if the two differ by 50,000 metres, you know something is going wrong

Like what your imprecise methods give you a very rough estimate where is the terminal and your legitimate customers in the nearby areas (not even doing anything for or against) are saying thanks for their shut off terminals.

https://upload.wikimedia.org/wikipedia/commons/b/b4/Comparis...

https://www.omnicalculator.com/physics/light-year

So -- for the purposes of locating a base station that's connected to the starlink network -- the starlink network would rely on the base station reporting it's location? Hopefully not... It certainly doesn't need to just trust it when it has direct observation available too.

I'm not suggesting that whatever GPSish thing you'd be able to get out of a pure starlink protocol communication would give you super resolution, but it'd certainly be enough to identify "Crimea" vs "Indian Ocean".

And as far as base station -> constellation -- putting a GPS receiver in it is of course smart to speed that up, but I'm not sure even that's necessary or just helpful.

Starlink can operate in a degraded manner without GPS now

https://gpsjam.org/

I guess Starlink receive the encrypted GPS signal, and send the encryption back. If they can’t decode it they can probably block the receiver.

The public normally use the unencrypted one with lower accuracy. Can fake GPS, but I don’t think it would last for long.

Spoofing is a big topic in the GNSS world (which is the broader set of satellite constellations like GPS). There's a lot of stuff happening to prevent spoofing. One of those things is using digital signatures that you can use to authenticate the signal. Another thing is using signals from multiple satellite constellations and correlating those. And of course the way these things work you need to talk 3-5 satellites from each network at least to be able to reliably interpolate the position. Even older networks, like GPS, change over time as satellites get replaced and software is updated.

The advantage Starlink has is that there are a lot more satellites that are also in a much lower orbit. There has been some talk about using their satellites for positioning. I think they look at that as a bit of a low value market so far.

Reply on news.ycombinator.com