Galileo's Ground Segment Problem | Hacker News Reader

Galileo's Ground Segment Problem

berthub.eu

141 pointsliotier48 comments

There seems to be this idea that Galileo is too expensive. I was curious so I looked up the cost for the US version. I found this:

"it had cost $10–$12 billion to fj eld the system,and the uSAF estimated the cost annually of sustaining minimal GpS services at $400 million"

That was in the 90s, so with 90s technology but also in 90s dollars. Not sure if firm conclusions can be made, but it gives an indication.

Source: https://history.nasa.gov/sp4801-chapter17.pdf

This one is very interesting, even breaks down cost per type: https://www.cs.cmu.edu/~sensing-sensors/readings/GPS_History...

I'm just impressed that a private citizen (? -- I am not familiar with the guy publishing these reports) has taken on the interest and duty to monitor and characterize errors in a global satellite system. Just to think that that would be something you could do (and not be lulled into thinking that some big corporation has it all under control) and go about making that your pursuit.

FYI: Bert is a super impressive person, a very bright mind and also the founder of PowerDNS (one of the more successful DNS resolvers in this world).

A totally agree. I saw him give a very interesting presentation with the title 'DNA: The code of life' during SHA201. Has also has worked on software processing DNA sequences.

Went down that rabbit hole. Thanks.

Check out 'Gunthers Space Page' https://space.skyrocket.de/index.html for similar awesome space knowledge.

Galileo is a European "me too"-system. As long as it seems to the general public to be running and problems do not make headlines, it is fulfilling its purpose. That the service is actually degraded and only working because of kludges like keeping prototypes running production is fine. It is the threat of maybe being able to replace Navstar or Glonass someday if really really necessary that counts.

I don't like to call these projects "me too", as it makes it sounds like the goal is only to brag about having one. It isn't. The goal is to guarantee independent access to a GNSS. It only needs to be good enough for US to lose the ability to use GPS access as leverage. I would refine that claim even further: it only needs to be good enough for Merkel, Macron, etc. to be able to confidently say "no" to a US demand made under threat of cutting GPS military access.

That said the Galileo project really looks like it is very badly managed, which is sadly rather common for European projects of this kind, especially on the military side of things. Even smaller projects with only a handful of countries tend to have so much overhead that it makes you wonder if going alone would not be cheaper...

> That said the Galileo project really looks like it is very badly managed, which is sadly rather common for European projects of this kind, especially on the military side of things.

Have you heard about the new system the UK was building to replace Galileo? https://www.theguardian.com/science/2020/jun/26/satellite-ex...

By comparison to the UK, the EU GNSS is a paragon of technical expertise and efficient project management.

The US still has that leverage because of their heavy handed way of treating even allies. The EU still very much relies on the US but not because they are friends and allies, and not because the EU could not do better, but because they have to make choices under duress.

Examples for such behavior abound but people judge with their feelings rather than the facts because EU citizens would rather assume they're in a stronger position, and US citizens would rather assume they're in a morally superior position.

> Under US pressure the EU has agreed to use transmission frequencies that could be easily disturbed or completely jammed by the US military. [0]

> The US has already leaned hard on European officials to abandon the €1.1bn (£772m) project. [1]

[0] https://tech.slashdot.org/story/03/12/01/1226207/galileo-sys...

[1] https://www.theguardian.com/uk/2003/dec/08/world.internation...

The biggest fear of US-Geopolitics basically since 1920 is close ties of europe to russia. A european economy fueled by cheap russian resources would mean even less reliance on the US. Thats why the US is so heavy handed to a lot of those projects. Fewer and fewer levers remain to nudge their "allies" into line, and that scares a lot of US-officials.

I'm not disputing any of that, the reasoning from the US perspective is pretty solid. US's real strength isn't their nuclear arsenal, it's the legitimacy its allies grant them, with the EU being at the top of the list. So the allies must stay on their side even if it means using these methods. From the US perspective losing the EU as allies is a double whammy because it also likely means US enemies would gain some, hence treating them as potential "enemies to be".

I also agree that fewer levers remain and we can see this with all the EU's recent political, legal, or economic moves that no longer really fall in line with US policy.

>> Under US pressure the EU has agreed to use transmission frequencies that could be easily disturbed or completely jammed by the US military. [0]

It appears "easily disturbed" means "on a different frequency."

According to ESA Galileo satellites have a different orbit which improves accuracy in the high latitudes of Northern Europe when compared to GPS.

Also the system is designed to be a primary reference, not just a GPS backup, and European taxpayers have paid €10 Billion for it. For that much money it should work correctly.

https://www.esa.int/Applications/Navigation/Galileo/Why_Euro...

They have two different atomic clocks too, one of which was supposed to bring accuracy down to 1cm or so.

The real question is why are "me too"-systems a problem?

To which I would opine: because the contractors working on them, and the layers of management managing them, and the funders funding them are well aware they're "me too"-systems.

No offense intended to the people working very hard on Galileo (of whom there are no doubt many), but it's hard enough getting contractors to give a $&@+ about actually critical systems.

If you tip your hand and let them know it's not really super important, is anyone surprised you get half-hearted effort and slipshod results?

But it's not really a "me too" system - if you take geopolitics into account, it's very much critical infrastructure. Beyond securing uninterrupted access to GNSS for Europe and taking away US's leverage, it also grants EU a seat at the table for some decisions the US would rather take unilaterally. With a working GPS alternative, the next time US wants to invade someone, they won't be able to cut off their GNSS capacity unless they get the EU (and, I guess, Russia) to cooperate.

That said, I'm yet to hear of a large contractor that cares, no matter how critical a project. Private market to me feels, in general, like the school system: with standardized testing and modern curricula, we get consistent but mediocre results, but we achieve that by filtering out anyone who cares about anything other than gaming the system.

Let's be clear, it's essentially a redundant backup system, a contingency plan. It is needed for the (unlikely but possible) scenario if GPS becomes hostile/unaccessible to EU for some political reason, and to prevent its access to be used as leverage for something.

It does not need to be "hot failover" - such things don't change that quickly. You might put up a few more dishes and satellites if that need actually becomes imminent in some future, but as you say, for now it's fulfilling its purpose.

But I'm a situation where GPS or Glonass would be considered 'hostile', there would also be the possibility of them being shot down which is not possible to defend against.

Galileo is also for trade reasons, so that reduced access can't be threatened. And it's also a military tool, after all GPS had been on reduced accuracy for a long time and this could happen again. And keeping an active technology development of such a vital infrastructure in house.

And for the consumer it's all great as most receivers can combine all these networks into improved accuracy.

> But I'm a situation where GPS or Glonass would be considered 'hostile', there would also be the possibility of them being shot down which is not possible to defend against.

Sounds like "the hostiles" wouldn't actually need orbital shoot down capability, they'd just need to knock out a bunch of those 20 ground stations. A team of guys with consumer drones and a half a kg of C4 each ought to be enough I'd guess. That's only "criminal gang" grade opponents, not even close to state level. (Or, you know, a bunch of random conspiracy nut QAnons or ProudBoys wound up on social media to believe the ground stations cause Covid or are part of a Jewish global banking conspiracy...)

In a "hot" situation, the ground segment is, well, "vulnerable" doesn't quite cut it. EU power projection isn't vaguely close to enough to protect all those far-flung islands that make the system work.

But there are "cold" scenarios where bombing ground stations wouldn't be kosher but denying GNSS services might be. That's what Galileo insures against. I don't think those scenarios are particularly likely, so it's awful pricey insurance, but it's also a very important capability so opinions may differ.

There the somewhat absurd ground station locations like Troll and Kerguelen helps. No one can go there on a whim.

Good point! But I suppose those uplinks could be hardened in times of war.

Yeah already the fact that support is built into receivers and that the frequencies are staked are very helpful.

After this last years of erratic US politics I think it was a good idea to try to gain more independence.

Were you thinking about Syria specifically or did you mean Iraq, or perhaps Afghanistan? Or the The Hague invasion act ? [1]

[1] https://en.wikipedia.org/wiki/American_Service-Members%27_Pr...

The "The Hague" invasion act is a openly hostile piece of legislature that I'm quite upset about. Every American should be ashamed of it.

Thanks. I am quite upset about it, but in aggregate the Dutch population doesn't give a <insert your favorite here>.

Threat management has cost, its capex (new sats) and opex (launch, integration, management). There has to be a huge upside, to justify more than a minimum.

I sometimes wonder if there is a model of GPS-like behaviour from the LEO objects, to get to something similar, but clearly not the same.

To answer your question, yes.

https://arxiv.org/abs/2009.12334

“This paper shows that continuous assured PNT (positioning, navigation, timing) service over ±60° latitude (covering 99.8% of the world's population) with positioning performance exceeding traditional GNSS pseudoranging would cost less than 0.8% of downlink capacity for the largest of the new constellations, SpaceX's Starlink.“

The risk analysis here, is that all the LEO stuff is US. So the same reasoning which took EU to galileo takes the EU to "what about LEO" but as an individual, in OZ, I shrug: none of this service comes from Australian owned satellites and I stand at risk of service disruption irrespective of who operates.

The eventual wider public ownership of satcomms, worldwide consortia, is a continuing story. Every nation state owns the vertical slice of air above its borders, they all have to talk about what happens in RF and satcomms orbiting through.

> Every nation state owns the vertical slice of air above its borders

only to 100km altitude. above that it's international space, pretty much how the middle of the atlantic is international space - free for all that manage to get there

Well, we don't hear much about the problems with GPS (though there has been some)

And in this case the problem was more geological in nature.

I can see why it would take 5 meter dishes to uplink to the satellites... but if you just need to measure the positions of the satellites, can't it be crowdsourced? If you have a few dozen people with fixed ground stations, rubidium clocks, and persistent internet, they could help out quite a bit, couldn't they?

You need to measure the position and then send that position to the satellite (as well as any necessary corrections to the clock).

One ground station can't measure position of a satellite, it takes at least 4 in 3 dimensions, all with synched clocks. I assumed (perhaps incorrectly) that knowing the true positions of the satellites was a problem due to lack of receivers, not really a problem of uploading.

Doesn't this non constant connectivity also have an impact on the SAR receivers built into them? I'd imagine those need a connection to relay emergency messages?

The MEOSAR network includes GPS, GLONASS and Galileo; plus the Chinese are supposed to be supporting with Baidou at some point.

I doubt the occasional downtime on some of the Galileo birds would make a large difference.

In any case, I think the MEOLUT ground segment for MEOSAR is distinct from the GNSS ground segment.

Thanks for explaining, I didn't know this! I thought it was a feature built into only Galileo.

Still, I believe this Galileo’s 5 meters error given in example is the best precision offered by the GPS?

I’m not saying this is acceptable, though.

2-3m accuracy is "typical" for cell phones. It's definitely not the best precision offered by GPS though. GPS receivers can listen to two frequencies, which allows for on-device correction for ionospheric error (this requires more hardware and processing, so it's not generally done on consumer grade hardware).

Surprised by the brexit comment. If true that seems like quite a petty move.

It does seem plausible though - UK moves on the space front have become increasingly erratic lately.

The UK enthusiastically backed the idea of a closed shop, where non EU entities would not be allowed to work on this project.

The idea was to limit competition against the UK's space sector, which is quite large.

This ofcourse means that now they have left the union, they are not allowed to participate.

This is entirely EU internal politics, from what I can tell - the UK has nothing to do with it. One of the author's earlier posts has some discussion about it: https://berthub.eu/articles/posts/state-of-galileo-and-accid... To cut a long story short, the satellites themselves and probably also the physical ground station infrastructure are provided by the ESA, which is not actually an EU institution and has non-EU members. Ther ground operations, however, are contracted out by the EU to various companies spread across the core EU states. The EU wants to take over European space activities for the same reasons they want everything else in Europe to fall under their control, so they're probably not on the best of terms with the ESA right now. I think there was also more political wrangling and changes of control that weren't covered in that post.

Possibly, what you're thinking of is the EU deciding that the UK won't have access to Galileo military positioning after leaving and pointing to the fact that the UK didn't want other non-EU states to have access to it as justification. Which of course they can do - it's just possibly a little counterproductive, given that the UK is still the big military power in the region due to the lack of willingness for anyone else to take that on.

It puts into context why the UK have been funding Atomic Interferometry research recently. If AI works, the likes of GPS, GNSS, Galileo etc. become obsolete because then we have GPS level accurate accelerometers which don't rely on satellite infrastructure at all

You're assuming the UK has a coherent policy that stems from brexit. I would dearly love that this is part of a grand unified 25 year plan (or in fact any plan that isn't "get boris to PM and keep him there")

The UK isn't a singular mind of Dominic Cummings. Plenty of institutional individuals and experts across multiple domains and departments have their own insights and directions to pursue.

But plans of mice and men rarely survive the heat of battle and political incompetence. But they try none the less.

> It puts into context why the UK have been funding Atomic Interferometry research recently. If AI works, the likes of GPS, GNSS, Galileo etc. become obsolete because then we have GPS level accurate accelerometers which don't rely on satellite infrastructure at all

I don't know about Atomic Interferometry specifically, but to me accelerometers -> inertial navigation -> intermittent need for an accurate external location fix -> GNSS not obsolete.

Reply on news.ycombinator.com