I'm more familiar with the older US GPS system than I am with Galileo, but I can make an informed guess.
GPS satellites broadcast "ephemerides" letting receivers work out where the satellite is - and some other details like clock error and predicted effects of the ionosphere.
You can make a rough prediction of these weeks in advance, but not precisely enough to meet GPS's precision requirements. Up-to-date short term predictions are needed to make things precise enough, so a ground-based 'ground segment' or 'control segment' broadcasts updates to these every 30 minutes or so.
The 'ground segment' is comprised of ~20 monitoring stations (at known locations, measuring the errors on signals from satellites), one 'master control station' that aggregates this data and decides what update to send next, 4 ground antennas that can send that update to the satellite no matter what side of the globe it happens to be on, and a backup MCS.
This is what is meant by "ground infrastructure".
The design of GPS is informed by its cold-war-era-military origins, so there's a backup master control station, and the satellites have an 'Autonav mode' letting them operate for 60 days in even if the entire ground segment has been destroyed.
Assuming the Galileo system has a broadly similar design, you can imagine the criteria for a four-day outage:
* Failure of both the master control station and the backup, in a way that takes more than 4 days to recover from. Maybe their backup control station wasn't fully commissioned yet, wasn't fully redundant, or some IT problem hit both sites at once?
* Discovery of some fault severe enough they felt they had to shut down until it was fixed, and the fix couldn't be rushed out.
* Either satellites have no autonav mode, or it somehow failed (such as due to bad data from the ground segment), or it's imprecise enough they decided pushing people onto GPS was a more responsible option.
The rumour I saw was that the entire ground segment is dependent on a single Precise Timing Facility in Italy, which went down and took everything with it. Not sure what the holdover capability is like without the ground stations, but I'd venture a guess it's not great,
According to a report from 2009 [1] "The Galileo ground segment has two GCCs, each with a Precise Timing Facility (PTF) providing a very stable reference GST to all elements in the GCC. The two PTFs are developed by separate companies; this diversification is applied to reduce the possibility of common failures in this important element."
I suppose it's possible the second GCC hadn't been fully commissioned? Or the switchover failed in some very hard to recover way?
According to the document linked, each GCC has its own PTF comprised of four Caesium atomic clocks. So between the two sites, they have eight atomic clocks.
Presumably if one atomic clock seems suspect, they simply look to their other seven?
Since TFA reports two sats being in `TESTING` mode (whatever that is), could this be that they have a fix that they are testing on those two satellites before deploying to the others? Ie the problem is on the satellite and not ground-based?
edit: no, the below link states clearly this is grounds-based trouble.
> technical incident related to its ground infrastructure
Given that the project is funded by €10 billion of EU taxes, I hope we'll get transparency sooner rather than later on exactly what's gone wrong, regardless of who it might embarrass.
No, that gives almost zero information as to the known cause:
> A technical incident in the Galileo ground infrastructure is affecting the functioning of the Galileo system, as a result of which there is a temporary interruption of some of the Galileo initial services.
> The cause of the technical incident is identified and recovery actions are implemented to ensure that the nominal service is resumed as soon as possible while safeguarding quality of the services.
"Galileo provides ‘initial services’ since December 2016. During this initial "pilot" phase preceding the ‘full operational services’ phase, Galileo signals are used in combination with other satellite navigation systems, which allows for the detection of technical issues before the system becomes fully operational."
Considering they are running this as an early preview I'm not sure if we are going to get much in terms of public response. Maybe now that it's a long outage journalists will be poking enough.
Pretty sure universities and other institutions would be able to inquire though.
Except if you want people not to be able to use the GPS signal, which sure make sense in military context (and even then, it's pretty meaningless, except if you change the key pretty regularly and you have a safe way to propagate it that won't be possible for enemies to get their hands on).
You know right that GPS signal is a one way signal? The receiver doesn't transmit anything, he just receive and use the timing and its data with some fancy math to be able find your position.
In a military context, they do propagate the key via their own, robust infrastructure. Over time, they've (USG) released keys and generally opened it up to the public, but they reserve the right to lock it down again.
As a result, while we have sub-5ft accuracy now, there's no guarantee that in war time, we won't lose that.
Is this correct? SA, or Selective Availability (an intentional degradation of accuracy) was turned off in May of 2000. This allowed existing civilian C/A code GPS receivers to generate better solutions immediately. It did not involve any keys or release of keys. I've never seen any "released keys" and have never heard of this. When did this happen?
Encrypted GPS do yu even make sense, what benefit an encrypted GPS will provide. Eu will know where were you when you requested something and even if your data was encrypted and i don't think you can intercept gps coordinates and you have to intercept at least three for it to be useful.
> Eu will know where were you when you requested something
You don't request a positioning signal, they are broadcast by the satellites and nothing is sent back by the receivers. Some devices (like smartphones) may send their GPS discovered position over the Internet, where encryption may be used, but this is not what is being referred to.
Encryption allows you to restrict access, or provided tiered access to signals of varying accuracy. You can also use cryptography in order to prevent spoofing.
I agree, although I'm not sure whether it's ever been implemented for public use.
You get anti-spoofing naturally if you're just using secret symmetric encryption (if your adversary doesn't know the key, they can't generate a valid signal to trick you with). To prevent spoofing against the public, you'd need to make use of public-key cryptography.
Yes, encrypted GPS makes sense. You can broadcast additional information that can be used for more precision only on encrypted channels, and have people pay for the decryption keys. (That's part of how the better-quality GPS for US military works)
Another way is selling online access to better-quality correction data.
Neither necessarily involves telling anyone where you are.
The encrypted channel just involves knowing a secret "key". This is in principle viable for high value military work, and you're right that Galileo has reserved such a channel.
In practice the Americans learned that COTS dominates for electronics. You can buy 100 of the special military product with the secret key, delivery in 18 months or for the same price you can have 5000 of the generic civilian product, 10 days lead time.
So these secret extra channels are of dubious value. But they're cheap to implement so they remain something new GPS style systems have.
In theory keeping the key secret defends against a sophisticated adversary spoofing your signal. But there's an obvious trade, if you only make ten units you can probably ensure the adversary doesn't capture one, but it's not very useful. If you make a million units that's initially more useful but your adversary will definitely get the key.
Even the US military doesn't use the encryption channels much anymore. Turns out that differential GPS is more accurate and you don't even need the encryption channel to do it. We already know the next message the GPS will send with high accuracy, the important part is detecting exactly when the first bit arrives at the antenna you everything you need from that.
If you want to get more accurate you don't use the encrypted channel you put a GPS in a fixed location on earth and have it send corrections to you. (called RTK)
Edit: Note that I have no information on what the US military uses their classified channel for. It is unlikely they get any accuracy information from it anymore given what civilian GPS systems have figured out.
Not trying to be funny here, but I would not be surprised if an automated Windows update borked something. Recently heard about an Antarctic base that lost heating for 36 hears after a Windows Update triggered a several GB download over the base's modem downlink.
> Not trying to be funny here, but I would not be surprised if an automated Windows update borked something.
If someone is running even important tasks that must not be interrupted, much less critical infrastructure like this, on consumer versions of Windows they're absolute idiots and should never be trusted with those sorts of decisions again.
If one insists on using Windows for uninterruptible tasks they should be using a server edition.
That said I'd always recommend
> Recently heard about an Antarctic base that lost heating for 36 hears after a Windows Update triggered a several GB download over the base's modem downlink.
That seems implausible, or at least to have been miscommunicated. The download and install phases are separate, so nothing would have gone down until the download completed. If the heating system required constant communication and went down due to not enough bandwidth being left over that's obviously a very poor design in general, especially in Antarctica. Even if that were somehow the case, it wouldn't take 36 hours for their network team to block/throttle the update traffic to let it work again.
The relevant point is whether it's set to do so automatically by default (as the consumer version is). Any operating system can be configured to automatically update, but many of them do not.
> The heater required an active internet connection and available bandwidth?
My guess is that the heater was being run by the machine that required the update and the update download process meant the machine was not available for usual operation. But I don't really know.
... unless it needs to do a quick follow-up update after applying one of the pre-downloaded ones. And then not three reboots later you get your machine back.
> Recently heard about an Antarctic base that lost heating for 36 hears after a Windows Update triggered a several GB download over the base's modem downlink.
I've been researching this. There was an outage. It wasn't caused by Windows. What is your source for this claim?
Assuming they're talking about the 2014 outage at Halley Research Station:
> Halley lost power on 30 July after the station’s generators overheated and shut down after a large coolant leak from the main arterial pipe in the station’s heating system.
The satellites themselves are most likely running some sort of proprietary or semi-proprietary RTOS for their C&DH computer (eg: VxWorks, QNX, Integrity, RTEMS...).
However, it's more than likely that there's a bunch of Windows machines on the ground segment side of the operation, since a lot of satellite fleet management software is Windows based.
Everything on orbit and on the ground is production product.
When you are several billion euros in, have 24 satellites on orbit, and have been broadcasting with the 'Valid' bit set for 2 years you aren't really in 'beta' as we use it.
They have been in production mode in all but name. Given the life/safety implications, there are some rather rigid qualification checks that have to be done to be 'fully certified'. That's what's been going on.
> In a statement published after this article's publication, the GSA blamed the Galileo outage on "a technical incident related to its ground infrastructure."
Guenter Hein, Professor Emeritus of Excellence at the University FAF Munich told us, “As far as I know, it is a problem of the PTF [Precise Timing Facility] in Italy – time has an impact on the whole constellation!”
They do (that's literally how GPS/GNSS works), but they drift, for various reasons. One of the parameters that GNSS networks broadcast is a clock correction (how far off each individual satellite's clock is, how fast it's changing, and how fast how fast it's changing is changing), and those parameters come from the ground stations.
It seems like a little drift shouldn't make that much difference, but keep in mind that the speed of light in a vacuum (or air) is about 1 foot per nanosecond, so for every nanosecond one of those clocks has drifted, that's a foot of positional accuracy you no longer have...
Nominally, yes, but the positioning error (according to the spec) is expected to be up to 425m after 14 days, growing worse over time until it hits a positioning error of ~10km at 180 days.
(later generation satellites have an 'autonav' mode that presumably makes this less severe (by using SV-to-SV ranging and such so they're not completely blind to changes), but Galileo may not have that capability or there might not be enough SVs to use that capability yet)
Wouldn't be too surprised if that ground station's job (which I guess is compensating for relativistic effects to the clocks' timing) was extended to compensate for rebooting clocks.
They do, but they’re less accurate than the ground station’s clocks. If not updated regularly, they’ll drift, with a gradual loss of positional accuracy.
we had this issues on our EC2 servers getting out of sync with other systems. Had to schedule some crontabs to ntp update yo. Maybe them EU dudes forgot to sync their clocks?
Sounds like a bug we hit at AWS a while back. The clocks on several of our instances started running at incorrect speeds. One was gaining something like 10 seconds every minute or something ridiculous like that. The rate was steady, just very far off from one minute per minute.
The Linux timekeeping system can normally deal with a clock that is steady but not the right rate by learning what correction to apply, but these were far enough off to be well beyond the maximum correction.
This only affected particular instance types, so changing to a non-affect type could fix it, but all the working types were more expensive than the types we were using which was annoying. Sometimes you could change to a working type then back to the buggy type and end up with one that did not have the problem.
There were a few people asking about this in the support forums, but apparently it was only hitting cheapskates like us who had not paid extra for support. Finally, someone paid for incident support and opened a ticket and Amazon quickly found and fixed the problem.
that was my first thought as well but surely they can replace it faster than this? also, the search and rescue signal still works. search and rescue signal are in a different band, so maybe antenna issues, or it has its own time reference. or maybe it's being jammed like the gps signals in the middle East, they are the same band.
A COSPAS SARSAT MEOSAR is essentially a separate platform that just happens to be traveling on the same vehicle.
As deployed today the two needn't have any electronic connection although they do.
Eventually a MEOSAR is intended to use the accompanying navigation transmitter to send replies from ground SAR. Initially an acknowledgement ("We see your distress beacon" and by implication "somebody will try to help") and perhaps later more.
But that requires proof of concept and then newer beacons as today's beacons, even the ones with GPS (many rely on COSPAS SARSAT to triangulate their position itself) do not know such an acknowledgment might arrive or how to interpret it.
All modern digital radio networks (cellular, police/fire, etc.) rely on GPS signals to discipline high-stability frequency oscillators. If you look at any cellular antenna you can always see the white mushroom-shaped GPS antennas, usually lower on the tower or sometimes just sticking up off the equipment building. Those oscillators are used as references to generate the transmit/receive frequencies and also for very tight timing between separate transmit locations (many digital networks actually broadcast the same signal on the same frequency from multiple places at the same time).
Without GPS, these reference oscillators can free-run for a while but they will eventually drift out of tolerance.
Basically, without GPS, most/all cellular/digital radio networks will fail in some manner after a while. Precise failure modes will depend on a lot of things including exactly what is wrong with GPS and the type of emission, how the base station software is written, etc.
This is true for S-CDMA networks (a.k.a. CDMA2000, primarily used in the US), but definitely not for the internationally more common asynchronous CDMA based UMTS (a.k.a. 3G or WCDMA) systems.
I'm not too sure about LTE, but some initial googling suggests that while it does need a precision frequency and time source, there are ways of deriving that from the network, which makes sense if you think about indoor base stations that often don't have a clear view of the sky.
Do they discipline by essentially being a full GPS receiver that decodes the timing information from multiple satellites, derives the time, and then adjusts the local oscillator so it is running at the right rate based on that time?
Or do they do something simpler, and just derive a frequency reference from the frequency of the carrier signals and sync the local to some fixed ratio of the carrier frequency?
The former would be more accurate, but would break if something went wrong with the data. The latter would probably not be as accurate, due to Doppler effect (I'm getting +/-0.00041% shift from Doppler effect), but would keep working as long as the carrier is being broadcast. You could probably correct for a good part of the Doppler effect just from the last known orbital parameters of the satellite--even if the data updates have been broken for a long time that should work.
It's the first - the GPS modules used for these purposes decode the time and then generate a digital pulse on an output pin at the top of each second (this is called a 1PPS output). A hardware counter counts the cycles from a local oscillator between pulses. The count between pulses is latched (this is all in hardware so far) and then the software reads the latched value. If your local reference oscillator is supposed to be 10MHz, your counter should report 10,000,000 ticks exactly. If it's high or low, the software can adjust the temperature of a little mini-oven that the oscillator resides in to speed it up or slow it down (some oscillators just use a voltage to speed up or slow down instead of an oven). Usually this is done in a PID loop or similar feedback loop. At this stage you can apply some filtering logic to help ride out certain errors (if you're suddenly off by 50% something else is probably wrong, etc.).
To ensure the GPS timing is as accurate as possible, you can have the GPS module survey its position for a lengthy period of time (since it is presumably mounted on an antenna mast somewhere and not moving) and average it out to establish a more accurate well-known location. The GPS can then use that known location to apply a correction to the GPS signals. Basically, if it's currently-computed position based on GPS signals is 50 meters away from the true well-known location of the receiver, it can figure out how far off the timing computation is and adjust. This is mostly handled by the GPS module itself, and is probably even more automated and accurate now than when I worked on it ~15 years ago.
There's also ways to increase GPS accuracy using supplementary broadcasts that contain localized offset correction information, referred to as Differential GPS (DGPS) run by many groups including the US Coast Guard and a similar technology called WAAS (wide area augmentation system) run by the US FAA. The GPS we were using didn't have the capability to use those sources but modern units probably do.
GPS-disciplined oscillators are based around full GPS receivers that decode the GPS signals and compute time from them in the usual way, then use that to steer the local oscillator. I think the better ones have a fixed-position mode which makes use of the fact the ground station doesn't move to get slightly more accurate timing, but otherwise they're basically the same as any other GPS receiver.
Not sure why you were downvoted. Our infrastructure worldwide relies a lot on location data, incl. GPS data.
There is actually an article from the New Yorker answering your question: much less traffic, all planes grounded, no train, etc. [1]
> The U.S. Department of Homeland Security classifies sixteen infrastructure sectors—including dams, agriculture, health care, emergency services, and information technology—as critical, and therefore particularly vulnerable to sabotage. All but three require G.P.S. for essential functions.
At least the plane stuff is wrong. Planes do not need to be grounded and will not be, but the capacity of the system will be reduced as we once again crowd into the rigidly-defined airways between the VORs. Approaches will use ILS as many do already and it will be all ok. Airplanes were ok for decades before GPS and even now can function without it.
this will be a fun thread!
high frequency trading relies on gps disciplined osscilators for the precise timing needed (ntp isn't good enough, ptp isn't even good enough, though White rabbit should be). so hft would cease or go wonky
trucking companies rely on it to enforce rest periods. so more truck accidents?
turn by turn direction on phones would go out. wi fi is enough for nav, unless your phone and the wifi are on different ntp networks maybe I had a phone with broken gps and I didn't really notice.
American trucker here. Occasionally I will lose the GPS signal in my truck, but my hours of service still counts down. The electronic logs are kept onboard, and transmitted at some point to headquarters.
I suppose the Qualcomm unit might eventually be off by a second or three. But I don't know if its clock is updated by GPS, phone or wifi. I'm guessing phone.
We're also required to carry paper log sheets in case of Qualcomm outage. So then it would be paper, watch and honesty. The honesty can be sanity checked by speed an distance.
I don't know much about Qualcomm's gear for truckers, but FWIW even if the machine does pull its time data over the cellular network (possible on CDMA, less likely - impossible? - on GSM), the cell tower itself likely relies on GPS for its own time sync. Its very common for GPS to be used as a globally-accurate source of time data even in applications that have nothing to do with location.
GSM also provides a time reference. I seem to recall connecting to some networks that didn't provide it though.
Base stations should be using GPS (and others) to discipline a local oscilator that should be ok enough for a while in absence of signal. They can probably also get a timing signal from their backhaul.
It works better with precise timing, but I think it's workable with best efforts.
Datacenters use GPS as a timekeeping source to keep synchronization across multiple locations more closely (or did who knows what Google actually uses, I could believe they wait till they move to a better solution to release their whitepapers).
> I had a phone with broken gps and I didn't really notice.
I certainly noticed in the UK on an iPhone 6 where the GPS chip was borked. You would drive for 10 minutes and then you would get an updated location. This was in outer London.
Galileo hasn't reached full operating capability yet, so that's probably why few people have noticed (as why would you depend on an incomplete system?)
Internet meme turned (potential) real life flash mob.
"Everyone go to Area 51 -- they can't stop all of us" essentially.
As to if it's just internet cranks, the US military shifting views, or aggressive, foreign AGI-PROP is debatable. My money says it's a distraction, but I don't know what it's a distraction from.
Can you imagine if these people decided to, say, storm immigrant detention centers, or abusive corporations... or anything that productive in a meaningful sense?
I had someone come and measure boundaries for some land on Friday, supposedly with centimetre precision. Does this mean that the markers they installed are likely in the wrong place?
It depends... were there any markers already present? In a past life I worked as a surveyor, and if a marker was present(it looks like a silver-dollar on the ground, with a spike holding it in place, these can last 100+ years), we would use that to measure from. If, however, they set up what looked like a tripod with a oval-shaped dish on top, that's a GPS receiver, and could be affected by this outage.
No - they placed new markers (orange disks on long spikes) driven flush into the ground, but they were walking around using GPS to do this (with supposedly centimetre precision, which I understand is what GNSS is for). This is in Eastern EU.
No, those markers work by listening to the GPS signals for an extended period of time and maybe PPK or perhaps an alternative network that's not satellite based.
"GNSS" just means satellite positioning systems in general, and e.g. "GPS" and "Galileo" are examples of GNSS systems.
At least regular receivers (e.g. in phones) utilize signals from multiple satellite constellations simultaneously, not sure if this is the case here as well.
Yes, I'm not in land surveying myself, but my company builds georeferenced 3D models of buildings at high resolution and sub cm precision, and we never rely solely on GPS information. There's too many things that can go wrong with GPS to reliably fly a drone around a building for example.
Yes, exactly. They use the signal as received at a nearby marker previously surveyed with high accuracy to calculate a local 'correction' to the GPS/GNSS signal, and give that correction to the roving units.
This is a good reminder to buy and use satnav modules supporting multiple positioning systems and technologies: most of GLONASS, GPS, BDS, Galileo, IRNSS, SkyHook, etc.
Italy has a pretty big space-related industrial sector, and excellent academic standards in some fields. It might not produce geniuses with the same frequency of the past (Galvani, Volta, Marconi, Fermi and so on) but it's not some random backwater.
My god, what a time to be alive, when you can't make simple jokes anymore, because somebody somewhere always is offended and thinks there must be some racist agenda in place.
Even as a German with a total lack of humour (and an Italian motorcycle with strange quirks in the garage) I found this funny. Funnier than the Holocaust at least.
>> [Italy is] not some random backwater.
> And nobody really thinks that.
As an Italian expat, I can assure you this is unfortunately incorrect.
> you can't make simple jokes anymore
No, you can't make jokes based on lazy stereotypes that end up affecting the lives of millions. It's a different thing. It might have been fundamentally innocuous, back when populations didn't mix as much as we do today, but now it's a real problem. If I go to a job interview and the potential employer assumes Italian == lazy, I'm done for. That's the casual bigotry that these jokes end up entrenching.
Nothing runs on time at Consorzio Torino Time (CTT), the guys quoted previously, also "The PTF in Italy been coordinated by the Consorzio Torino Time (CTT), with the partnership and support of SpT and T4Science in Switzerland." so i'm pretty sure as an Italian who enjoys to stab Caesar and switch sides during war, THE BLAME IS ON THE SWISS
GPS satellites broadcast "ephemerides" letting receivers work out where the satellite is - and some other details like clock error and predicted effects of the ionosphere.
You can make a rough prediction of these weeks in advance, but not precisely enough to meet GPS's precision requirements. Up-to-date short term predictions are needed to make things precise enough, so a ground-based 'ground segment' or 'control segment' broadcasts updates to these every 30 minutes or so.
The 'ground segment' is comprised of ~20 monitoring stations (at known locations, measuring the errors on signals from satellites), one 'master control station' that aggregates this data and decides what update to send next, 4 ground antennas that can send that update to the satellite no matter what side of the globe it happens to be on, and a backup MCS.
This is what is meant by "ground infrastructure".
The design of GPS is informed by its cold-war-era-military origins, so there's a backup master control station, and the satellites have an 'Autonav mode' letting them operate for 60 days in even if the entire ground segment has been destroyed.
Assuming the Galileo system has a broadly similar design, you can imagine the criteria for a four-day outage:
* Failure of both the master control station and the backup, in a way that takes more than 4 days to recover from. Maybe their backup control station wasn't fully commissioned yet, wasn't fully redundant, or some IT problem hit both sites at once?
* Discovery of some fault severe enough they felt they had to shut down until it was fixed, and the fix couldn't be rushed out.
* Either satellites have no autonav mode, or it somehow failed (such as due to bad data from the ground segment), or it's imprecise enough they decided pushing people onto GPS was a more responsible option.