because we all know that no government would ever...
because we all know that no government would ever fuck with GPS signals, except for Russia, the USA, etc...
A recently published research paper proposes a system for terrestrial positioning that could give greater accuracy than the existing satellite-based systems, and could potentially be incorporated into future mobile networks. Global positioning systems like GPS, Glonass and Galileo are widely used for applications such as …
of course it is, because which shopping chain wouldn't want to know not only which shop you're in but which item you just picked up to look at?
I trust this, if installed, can be turned off as easily as a GPS chip.
(Feeling fairly grumpy about location at the moment; a new diabetes app 'we don't take any data, we don't share any data' will not allow a bluetooth connection between the supported blood test meter and a phone without location not only being turned on but with a valid position. I can see no reason why this is relevant and await the manufacturers response to my query).
Mobile phone permission systems require 'location' permission to be able to scan for bluetooth devices because bluetooth beacons can be used to build up a location profile and provide low-accuracy positioning of the user (see apple tags, but it works both ways).
I have no idea why it would need a valid location though.
Catch-22
Old Android (don't know about iOS) used to let you use Bluetooth without switching on location.
But I bet your diabetes app has been built with a copy of the SDK that - pointlessly - demands a newer version of the OS, preventing a downgrade (or just a cheap old phone to run that app).
As for requiring a valid location: probably the latest OS is getting stricter just to squeeze as much out of you as possible.
> But I bet your diabetes app has been built with a copy of the SDK that - pointlessly - demands a newer version of the OS
Pointlessly, like being allowed to actually publish the app? You must build with a current SDK (or within the last 1 - 2 years depending on Android / iOS) or you can't publish your app.
Pointless, yeah sure.
Yes, making that a requirement to publish - you think that *that* isn't pointless? So long as the app can work, so what if the SDK used didn't expose the latest set of OS functions?
Or do you seriously believe that backwards compatibility is a Bad Thing?
If so, it will really pain you to know that I've got programs originally built on Windows 2000 that still run on my Windows 10 box and are 100% as useful now as they have always been. Even the ones with GUIs.
This post has been deleted by its author
Yes, I can connect to generic serial-bluetooth adaptors and use them (e.g. with a bluetooth terminal program) without issue and without location services. I can connect to the blood sugar monitor, but even if it's already connected it still asks the OS. Very strange.
> await the manufacturers response to my query
This is a reaction that one sees quite often to the "OS ties BT and location together" issue - Amazon reviews for pretty much anything with BT are chock full of annoyed comments.
Looks like it is up to us to keep telling everyone that the manufacturers can do nothing about it, it is a problem baked into the 'phone OS and everyone should also be sending their complaints direct to Google (probably Apple as well - I have no iOS to check), plus BBC Watchdog, your local Privacy Activists etc. Complaining in public about things like this - one good thing that Twitter could be used for?
And stop buying things that use BT to talk to phones. Agreed, that is a problem if every damn diabetes test device uses BT instead of having its own *legible* display (wish Boots didn't seem so damn proud of those devices, putting them at the top of the display shelf).
Apple doesn't tie the Bluetooth and location permissions together, which could theoretically allow an app to use Bluetooth devices as the only method of tracking, but not as well. I generally prefer that to Android's method of making every app that interacts with a device have location either because they have to or because the app wants the data, and good luck finding out which it is. Still, what might be better is the OS allowing an app to request a Bluetooth connection and the OS goes and finds the device and connects it in. Instead of letting an app control the Bluetooth hardware directly if all it's going to do is open a serial port, the OS could do that and reserve the permission for apps that the user wants to have full control over the Bluetooth hardware.
I have exactly the same problem with the Canon printer app on Android. It will not do anything unless you allow it to know its location, and precisely at that! I assumed it is to find the nearby printer but it should be able to do that via WiFi. So the app is no longer on my phone.
That is not the fault of the diabetes app. That is the fault of Google or Apple, depending on which type of phone you own, and the way they designed that feature.
The manufacturer will be confused as to why you are questioning them about how Bluetooth works on your mobile device and point you back to Goople.
I'd actually say that it's the "fault" of the EU, and advertisers - something that can be used to identify your location should be switched off with the location control, which includes wifi and bluetooth scanning which could find nearby devices (with a known location) and locate you as a result.
Of course this could be mitigated easily if the devices instructed you to make the bluetooth connection through the OS before starting the app rather than trying to corral you through it.
> I'd actually say that it's the "fault" of the EU
WTF? How on earth is this the fault of the EU (except because of some rabid and irrational EU hatred?)?
It is Google who wants needs location data, to better sell your information, that's all. So they conceived of this new law of physics which states that Bluetooth needs location, never mind if it worked very well without it before. (Which will most likely be retconned soon to state that Bluetooth has always needed precise location data, by design.)
Legislators could outlaw it if they wanted to.
Given how helpful location data has been in determining who exactly was present within the geo-fence around the US Capitol building on Jan 6 2021 it is doubtful they'll ever limit location tracking.
> Legislators could outlaw it if they wanted to
You know that they don't *really* sit around all day making laws just for the fun of it, don't you?
Legislation is created in response to a perceived need; if there is no perceived need, everything is ticking along, why would the legislature bother? If you want to have a law to force BT unbundled from location then you can always act yourself - open a petition at least - make it clear that this is causing societal damage and that there is a *reason* to legislate against it.
Then you get commercial lobbyists involved and so on.
Completely wrong. It was discovered that Bluetooth access allows you to determine location in many cases.
The way it works makes "fixing" this impossible.
The only moral solution was to therefore add it to the location permission (though I'd have preferred they keep the coarse vs. finegrained location permissions separate)
If this "moral" solution was pushed onto Google by the EU, I'm all for it.
The only other option was continuing ignorance of location leakage.
You're right about that, but the OS could be designed with more granularity to prevent abuses. The problem is that an app given fine control of Bluetooth hardware can, in some but not all cases, determine location. Here are some options to help with this:
1. Provide more options for apps to do something with Bluetooth that don't require full control. For example, if they're going to connect to a single device, have the OS do the connection and just give the app the one connection to work with. That app will not have the ability to scan around for beacons and therefore couldn't track location.
2. Split the permission. Just because an app really needs full control of Bluetooth doesn't mean it should automatically have access to satellite navigation receivers too. One permission could be "Fine Bluetooth WARNING may leak your location", with the other permission separated. If I allow an app to use my Bluetooth hardware but I have no beacons in range, it won't know where I am. If I'm forced to let it have all location access, it will get that data no matter where I am.
Just because full control of Bluetooth hardware can leak location data doesn't mean we can't improve the situation.
Fair enough. And I totally agree - if a more restrictive API to bluetooth could be created that would allow most use cases to work correctly without giving away location information, then I'm all for it.
And you're right, if no beacons are in range, there shouldn't be an issue with you bluetoothing to a local device... To be honest, I hadn't really thought of the all the possible scenarios!
So, it does seem rather lazy to take the solution that they did - and I hope it's temporary - but it was far more responsible to do that than just leave it as it was, and so my reply was pointing out the paranoid absurdity of the comment I was replying to.
Unfortunately, they seem to be making permissions less fine grained - every new release seems to be more hobbled to make it "consumer friendly". It's an excuse that annoys me.
Notice how "internet access" is no longer a thing? (well, it is, but it's just granted without asking/notifying the user)
course/fine location seems to have gone. (EDIT: Apparently , it's still a thing https://developer.android.com/training/location/permissions#accuracy )
And I'm still peed off that there is no "startup manager" - I have alexa installed for when I'm over my mums - the thing isn't used otherwise, yet it just HAS to start up every time I power on my phone unless I force stop (which is temperamental) or delete it. That is not user friendly at all.
Now, I know, that apps that what to produce alerts need to make a callback to register with the service to work, hence why apps could "officially" need to startup, but
1) I often don't care about app notifications.
2) It's an easy way for apps to be abusive.
3) The design should be changed so that it's not necessary!
Still, I expect things to get worse, not better. I'm surprised the "developer" menu still exists!
Anyway, I've gone off on a tangent...
cheers!
Let's not let the manufacturers off the hook so easily. First, they should know the privacy implications of their app, even if it's OS/SDK related. Second, they should be pushing back on Apple/Google about tying disparate technologies, as well as violating the trust of consumers. I have always felt that Android security isn't granular enough. Apps that are not dependent on your location should have the option of turning off location services. The only time GPS should be required (ex, you can choose not to use an app) is for emergency services. And even there it's only necessary if you are in an area where your phone can't talk to multiple towers.
You should read this on the changes to android location security: https://proandroiddev.com/evolution-of-location-access-on-android-897a0449e71f
And also, as to granular location-access: https://developer.android.com/training/location/permissions#accuracy
"... will not allow a bluetooth connection between the supported blood test meter and a phone without location not only being turned on ..."
I find that annoying too (have come across it when using my phone to control sound equipment). One way around having to give permission for location is to use a tablet or something without GPS.
Unless there is a good reason for the app to connect to the Internet you can use a no root firewall and block it from doing so.
I was trying to use a Sonos a long while ago which same as yours required the location switched on for everything (including Bluetooth) set up. Not sure about usage as I didn’t get that far as you’ll see. It didn’t say that this was an OS requirement and I got the feeling it was their choice. I wasn’t keen about this anyway so I stuck it on a spare phone to be certain. I have a firewall on that and also no SIM card in it either.
I tried to run setup and the app didn’t like no internet or location, in fact it refused to work full stop. I couldn’t see why it needed internet access in the first place it’s a just a bleeding wireless speaker after all. I spoke to Sonos who were not vey good with explanations for these requirements. They were happy to tell me that it was only during set up but not able to explain why it needed it.
I didn’t own this kit it belonged to my then housemate who thought it was brilliant. Instead I bought a Bluetooth adapter and just used that to connect to my HiFI instead.
Because of Bluetooth beacons - the indoor positioning system. Unlike GPS, the beacons send only a GUID. An altruistic company, like Google or Apple, serves a database of their actual location.
This is my suspicion behind Google and Apple wanting the headphone jack gone.
Companies have been selling the kit, including eye-tracking, to shops for years. It's now fairly easy to combine wifi, Bluetooth and camera data for positioning and tracking. In some countries the mobile networks will also give you positional data, for a price. Not sure of the utility of such detailed tracking above and beyond existing placements of special offers, even if there have been reports of unexpected correlations: beer and nappies springs to mind but I can't recall the details or be bothered to search.
Having worked in mobile telecoms in the time of GSM and into the early part of this century. I can assure you that without very accurate timing mobiles do not work. Time of light propagation delays are built into the systems. Signal from three towers would even in those days give you a 100m diameter circle for a location. Modern systems much better resolution for obvious reasons.
Neil, I would send the kit back as unsuitable for the vast majority of homes. LibreFree2 does not need such stupidity.
(Anon because of the last sentence.)
Hmm, I seem to remember that mobile networks and digital TV depends on GPS (or system of choice) for correct operations, so at best this is to help navigation in dense areas and is in no way a usable alternative to GPS due to said dependency. Unless they totally rework it all to work without GPS and relying on optical timing and some other knowledge of the base station locations, all at huge cost and effort, which won't happen without a gov mandate and them being expected to pick up the cost...
Yup, various permutations of network signal time of arrival positioning are already in 3GPP.
They've been gradually ramping up the accuracy requirements over the years.
Starting from ~100m in GSM days, they've now got the 5G iteration to a best-case accuracy of 30cm.
(Should also note that Cambridge Positioning Systems had a way of getting down to ~10m around 20 years ago)
GPS used to be capable of decimeter precision. It was intentionally nerfed by the US government to prevent it from being used for weapons guidance after September 11, 2001. About 2004 as I recall. I used to have a Garmin GPS12 that was capable of MGRS (Military Grid Reference System) positioning to 10 digits (5x5) which equated to a 1 decimeter square area.
AFAIK, this is no longer the case, it was de fuzzed in 2000 according to this : https://www.gps.gov/systems/gps/performance/accuracy/
It's quite possible to get cm accuracy using RTK, I work on such a unit for phones, and that is publicly available technology.
During the Gulf war the US had to rely on civilian GPS in a hurry so they turned off the "selective availability" feature normally in place during peace time.
Yup, exactly the opposite of what you would expect (war = degraded accuracy). The Clinton administration realised this was pointless so changed policy:
https://www.gps.gov/systems/gps/modernization/sa/
It was intentionally nerfed by the US government to prevent it from being used for weapons guidance after September 11, 2001.
Right from the start it was designed to be non-functional at high speed & altitude, to prevent its use in missiles. Selective availability to reduce the precision of the location was added around the time of the first Gulf war in 1991 and was removed on May 1st 2000. 9/11 had nothing to do with it.
Which Galileo system do you think the U.K. is “missing out on”? The one that still hasn’t achieved Full Operational Capability, ten years late, despite having spent $15bn out of its original $4bn budget? It now looks like it *maybe never will*, despite billions of ongoing spend.They continue to launch satellites occasionally, but now they are coming to the end-of-life and failing in-orbit faster than they launch them.
The “great technical secret” of Galileo, is that while their webpage trumpets “needs 24 satellite plus two orbiting spares”, of which they have 23, that’s a very fundamental misunderstanding / misrepresentation of the orbiting constellation. What they actually need is three orbital planes each of [six satellites plus spare]. Satellites are launched in pairs, and can’t be moved between orbital planes. Among other things, that means if you have 6+5+5, you need two double-launches not one to reach six in each plane.
But also, the orbital spares: nobody really knows where the claim “needs *only* two spares” comes from. It’s gibberish. Probably. they planned to get to 6+6+6, without failure, launch another for luck to get to 8+6+6, and claim “done it tah dah”, which would basically be a lie there would be redundancy on two planes. Then very quietly, two more double-launches to reach 8+8+8, again assuming no orbital failures.
Their entire plan assumed zero launch failures, and zero failures in-orbit. In an industry where 10% launch failure is normal, and in-orbit lifetime of 12 years is defined at “95% probability per spacecraft”. If you have 25 spacecraft, you’d *expect* another in-orbit failure.
They’ve actually had three in-orbit failures inside lifetime. The oldest of the operating spacecraft is 11 years old, 12 year intended lifetime, so as from 2023, we expect them to start failing - about four per year (the rate they were launched at). EU planned launch cadence for the new ones is four per year, so once they fall below operational numbers there’s no way to recover. And they are now blocked from launching because of Ariane 6 delay until “late 2023” at best. The EU plan of using the “Arianespace” launches which were actually rebranded Soyuz, has been badly exposed by global events.
So, this is EU space program you think the U.K. missed out on. Failed to plan redundancy. Redefined its terms of success, then believed its own hype, getting badly caught because it’s own management didn’t understand basic orbital mechanics. Poor judgement on supplier selection. Bad constellation design. Likely never to deliver actual valid product. But continues to spend billions per year, and is forced to do so to the maintain the never-quite-there constellation until 2040+, because it can’t face the humiliation of cancelling it.
And worst of all, of course, in the 2035 timescale, nobody is going to be needing a *satellite* navigation system, because better technical solutions are being developed: atomic clocks and gyroscopes, allowing accurate dead reckoning inside smartphone without relying on externals. Everyone educated in the field knows it, but noooo the EU had to make a monkey-see-monkey-do copy of a legacy system just to prove they could.
LORAN could never have been accurate enough for many of the uses of GPS/Galileo/GLONASS etc for both positioning and timing. This look like a terrestrial-based system that could do that. However it would only really work well in urban areas where signals from several base stations are available. It would work even better if the mobile companies would all agree to hosting the common signal so receivers could use all networks. It would still be a problem in rural areas where some networks co-host base stations on the same mast so there aren't multiple independent pseudo-range vectors to feed into the multilateration algorithm.
Like satellite-based systems, the one proposed by the researchers relies on the accurate measurement of arrival times of radio signals, but it uses a time signal distributed to all the radio transmitters via an optical fiber network to ensure they are synchronized to a common reference clock.
Elsewhere the article mentions sub-nanosecond time synchronization support.
With this the researchers achieved position accuracy to the decimetre (why not '10cm' or 0.712857 linguine?). What level of accuracy could they achieve without the improved time synch? Which side of which street you're on? Sounds good enough for me.
combine harvesters for instance...
Yes I can see that they need accuracy. Also diggers on construction projects and many other things. I doubt, however that in the middle of a field or HS2 destruction site there will be 6 cell towers in range to triangulate from. My guess would be to drop in some temporary* beacons with known/calculated locations. Building something that can provide that level of accuracy for a phone does not seem to fit that use case. 'Which side of the bulldozer did you have your phone mounted?' ... 'Ah, that explains the collision'.
*Temporary = many years (depending on the project).
"accuracy to the decimetre (why not '10cm' or 0.712857 linguine?)."
It's not a commonly used unit, but decimetre is a valid unit. I remember the school rulers we had in primary school when the metric system was bring introduced in schools back in the day. There were various useful things engraved/printed onto the wooden rulers to be helpful, such as inches down one side, centimetres down the other, millimetres marked along the first 10cm and definitions of things like decimetre=10cm=100mm and dekametres=10meters (That would be decameter in USAian although why they would choose to have a different and unique local spelling of a metric unit in a non-metric society is beyond my understanding :-))
Sounds feasible. But does anyone really need it? I can't think of an application where differential GPS wouldn't provide all the accuracy one would need. But that may just mean that I have no idea where accurate position might actually be helpful. Surveying maybe? And I have no idea how well/badly DGPS works in the worst cases. The internet tells me how DGPS works (already knew that), but not as much as I'd like to know about practical limitations. I'd like to hear from anybody who has hands on experience with this stuff.
It seems to me that the range of this system is going to be limited. How many cell tower signals would one need to get an accurate position? (three?) In rural areas in North America you're often lucky to get one signal without stopping, maybe climbing part way a hill, and holding your phone just right.
I suspect that the amount of usable signal needed for an accurate time check from multiple sources would be far lower than the amount of signal needed to carry a phone call.
also a minimum of three sources would only be needed for an initial fix, once fixed a system could work on only two (since you are unlikely to move between the two possible positions in the time between readings).
Obviously the more sources of signal the better the fix accuracy will be, but when signals are lost most systems degrade gracefully either using dead reckoning or using inbuilt gyros to work out where they are moving, even the decade old pioneer stereo in my car has gyros and a compass built in, using these combined with a speed pulse from the ECU results in surprisingly accurate tracking when gps reception is blocked. although it did take the system a few 100 miles of driving for it to train itself for the sensors (presumably it needed to work out orientation and how far the vehicle travelled per pulse from the ecu)
"But does anyone really need it? I can't think of an application where differential GPS wouldn't provide all the accuracy one would need."
The obvious application is identifying and tracking the precise location of any person anywhere, even without satellite line-of-sight, even indoors or underground. Every phone becomes a continuous proximity reporter, generating very high resolution signal-field intersections with every device in range. Greater population density improves resolution. Claimed accuracy is approximately half the thickness of an adult's torso.
On the upside, finding a lost (device equipped) child is nearly instantaneous. On the downside, finding and tracking or potentially removing any member of the population is similarly instantaneous.
It's the institutionalization of Apple's AirTag problem.
"I can't think of an application where differential GPS wouldn't provide all the accuracy one would need."
Urban canyons, subways, parking ramps, Lower Wacker Drive in Chicago, and similar places where GPS signals have difficulty due to limited line of sight and/or multipath.
Sounds feasible. But does anyone really need it? I can't think of an application where differential GPS wouldn't provide all the accuracy one would need. But that may just mean that I have no idea where accurate position might actually be helpful. Surveying maybe?
Surveying for sure, construction, forensics (traffic accidents) among others I know of, but probably the biggest one that would get telecoms companies on board is AR apps. The others are somewhat specialised.
When you do require that level of accuracy, RTK(or DGPS) frequently doesn't deliver, particularly in built up areas with building reflections, bridges, indoors, also the antenna used by dedicated RTK systems, while it can fit in a dedicated unit, are simply not suitable for inclusion in a phone. But they are getting smaller and cheaper all the time, and it's pretty feasible to construct your own. For our use cases, it would be super handy to be able to say, oh, you don't need any extra equipment, just get this phone and our app.
"In rural areas in North America"
I got the impression from the article that this system is intended as an assist/replacement for where GPS etc doesn't work or can be outright wrong such as built up areas where the sat signals may not be visible to the device. Out in rural land, GPS etc is all you need unless you are in a narrow mountain valley or gorge.
The towers are in fixed positions, perfect for receiving GPS (and the other positioning satellites) signals since the correction for apparent positioning errors can also correct the time base.
They'd work very well as differential GPS sources within the GPS framework, there's no need to reinvent the wheel and create yet another positioning protocol.
"We can do GPS with cell towers" assumes there is a wonderful blanket coverage of nationwide cell service. GPS works everywhere. And those times it is most desperately needed are generally in the middle of nowhere, with nary a single bar of cell service.
On the other hand... If you want to cheaply track everything & everybody in a high-density urban setting, this is just the ticket for that population's control.
GPS is blocked by structures to a far larger extent than cellular, so there will be many occasions where you don't have GPS but do have cell service inside a building. There may be a cellular antenna inside a tunnel, or inside a subway where GPS most definitely does not work.
Now true if you are outside of cell range this doesn't help you, but the idea is to improve upon GPS by providing more accurate time and position not replace it. If you are lost in the mountains, you probably don't care if your position is off by a few meters or even a few tens of meters. If you (or your rescuers) can establish your position within that granularity that's plenty good enough. There is no need for a cell tower to get that down to cm level precision.
"GPS works everywhere. And those times it is most desperately needed are generally in the middle of nowhere, with nary a single bar of cell service."
There are mobile not-spots and GPS not-spots. In general, they don't overlap much other than indoors, a tunnel etc.
This is rather chicken and egg. Many moons ago telecoms systems used atomic clocks to feed synchronisation links to much of the equipment that required accurate timing. This was superseded by using GPS receivers as timing sources as cheaper and more flexible. At least some telcos still have their atomic clocks as backups for GPS but given the intricacies of network synchronisation and the record keeping associated with it even where the telco has backup systems i would not expect a clean change over in the event of GPS failure. A move back from GPS to telco provided synchronisation would require extensive network replanning.