11

GPS satellites don't transmit strong enough to reach indoors, through the roofs and walls of buildings, like cell phones do. GPS signals that enter buildings through windows are unreliable since they often have bounced and thus give the wrong distance measures by even hundreds of meters. Since cell phones work fine indoors, and GPS uses the same frequency, I suppose there's no need to use another frequency.

Would it suffice for today's kind of GPS satellites to orbit in low Earth orbits, such as planned communication satellites constellations like Starlink? Would it be feasible to equip them with larger solar arrays, or would nuclear generators be required?

One can imagine many commercial applications for precision indoor navigation (like more efficient robotic vacuum cleaners ;-) Just reaching through the roofs of single storage factories and shopping centers should make for a great market. The alternative today is to install radio or ultrasound beacons (or perhaps a camera system) in every room where one wants to have a location service. A single ubiquitous system that's already standard would have economic advantages.

DrSheldon
  • 47,946
  • 12
  • 162
  • 329
LocalFluff
  • 26,981
  • 8
  • 82
  • 222
  • 1
    Not all cell phones work fine indoors and not in every building. There are metall roofs and steel construction buildings. Multi path distribution of satellite signals does harm the GPS position precision. The speed of electric waves is different in air and in solid materials ( non metallic solids ). But GPS even handles different speed of light in vacuum and air. But how to handle the much lower speed in solids? – Uwe Apr 02 '19 at 11:20
  • 4
    All you have to do is construct your building appropriately. GPS works perfectly well inside my (wood frame, single story, non-metal roof) house. – jamesqf Apr 02 '19 at 15:56
  • @jamesqf Really? It's been some years since I disappointingly tried out GPS inside a hut, and maybe with already then old equipment. The receivers have been getting alot better. But how far could that development go? – LocalFluff Apr 02 '19 at 17:21
  • @jamesqf If the (wood frame, single story, non-metal roof) house has a thatched roof or a shingle roof soaked wet from a heavy and long rain, GPS receiption may be weak. If the roof is dry, GPS may work perfectly. – Uwe Apr 02 '19 at 20:25
  • @jamesqf: Unfortunately that interferes with other desirable characteristics, such as good thermal isolation. – MSalters Apr 02 '19 at 23:57
  • 1
    There is no need to install radio beacons for a indoor positioning system. Wifi access points which are ubiquitous these days can do that job. Wifi positioning system (https://en.wikipedia.org/wiki/Wi-Fi_positioning_system) is already used extensively by mobile phones and laptops and have a better accuracy than GPS indoors. – Kevin Selva Prasanna Apr 03 '19 at 08:36
  • 1
    The comparison with phones is really unhelpful. You yourself state why: "since they often have bounced". While that is a problem for GPS, your phone doesn't care at all, hence it works – Hobbamok Apr 03 '19 at 10:05
  • @Uwe: Granted, I've never tried using GPS indoors when the roof was wet, since I live where rain is not exactly commonplace. And I've only tried it a few times, mostly to see if GPS devices were actually working correctly. I mean, why would you want to use GPS indoors? Unless you've been kidnapped or something, you already know where you are :-) – jamesqf Apr 04 '19 at 04:42
  • There are dozen of different technologies that are currently being commercially used for indoor tracking. While Beacons using Bluetooth made their way in Airports and Museums the Industry (especially intralogistics) and Sport (e.g. NFL) is mainly leveraging Ultra-wideband (UWB) due to its physical features. See the following RTLS Technology comparison – Petr Passinger Apr 04 '19 at 11:09
  • @PetrPassinger Oh yeah, many different concepts indeed. From ultrasound to image recognition and whatnot. Which is the problem! Because the GPS receiver everyone already has doesn't pick it up. So only specialists use other systems. – LocalFluff Apr 06 '19 at 22:17

5 Answers5

14

There are some big differences between GPS and cell phone signals:

  1. GPS relies on the exact timing of a signal (1 nanosecond off = 30 cm of position inaccuracy). Cell phones are much more tolerant to variations in signal timing. Basically they don't care as long as the packets arrive quickly enough not to cause a gap in the audio.

  2. GPS transmitters are 18,000 km away, cell phone towers are less than 5 km away. So a cell phone signal is much stronger at the receiver.

item 2 could be solved with a stronger transmitter, but that would make #1 worse: with a stronger signal, you get more multipathing: more reflections that are still strong enough to be picked up by the receiver.

You can't really use indoors transmitters either. Many buildings have a steel structure which reflects radio waves, so you get lots of multipathing when the transmitter is inside.

Hobbes
  • 127,529
  • 4
  • 396
  • 565
  • 9
    "... with a stronger signal, you get more multipathing..." This sounds like something that's made up and not based in EE fact. Any digital radio receiver system will have automatic gain control (AGC) before the ADC. So as far as the digital signal is concerned, the multipathed reflections will have the same strength relative to the direct path signal no matter what the strength of the satellite's signal is. – uhoh Apr 02 '19 at 12:06
  • 4
    if your signal is weak to begin with, reflections can drop below the noise floor. Strong signal->more reflections make it to the receiver above the noise floor. – Hobbes Apr 02 '19 at 12:28
  • 4
    It doesn't work like that. It's not like there is a "floor" and signals below it disappear. Every signal (of a given type, e.g. L1) from every GPS satellite in the sky, plus every reflection of every signal from every satellite in the sky plus all sources of noise, all go through one single front end amplifier and then one ADC, where it is digitized into one digital stream of 1's and 0's. That stream is then copied to dozens of identical correlaters, each looking for a different Gold code from a different satellite. – uhoh Apr 02 '19 at 12:35
  • The one AGC before the one ADC will adjust the level so that the strongest carrier is at some predetermined level. At this point (before the ADC) the levels of the reflections will have the same proportionality to the direct signal, independent of how strong the constellation is broadcasting. I really don't see how multipathing can be worse for a strong constellation of satellites than for a weaker constellation of satellites. However, front-end noise fixed by the input FET will be a bigger problem for the weaker constellation. – uhoh Apr 02 '19 at 12:37
  • 1
    1, 2, 3 – uhoh Apr 02 '19 at 12:42
  • 2
    @uhoh, would need to find a reference, and is a analog artifact but was taught a potential solution for https://en.wikipedia.org/wiki/Ghosting_(television) where the cause was proximity to the transmitter allowing reflections to be above noise floor was adding an attenutor or just making the antenna less effective. Thinking through may have also been result of of doing AGC later in the signal path than GPS. – GremlinWranger Apr 02 '19 at 12:43
  • @GremlinWranger I see what you mean re later in the signal path than for GPS. It's also a different kind of modulation, there's more opportunities for nonlinear effects, there's no ADC followed by correlation and decoding. TV transmitters can be ~100 kW so being in close proximity might introduce other problems besides saturation. I would be happy to be wrong in all this because it means I'll learn something new. – uhoh Apr 02 '19 at 12:48
  • 2
    @uhoh, other place I've met power reduction to reduce multipathing is in Sonar, which also does not say anything useful at all about GPS and their somewhat unique receiver design so think I get to learn something today. – GremlinWranger Apr 02 '19 at 13:12
  • @GremlinWranger Thanks! There is a DSP SE site that could handle a question like this, but I'm not sure how well a question of this multidisciplinary nature would be received. – uhoh Apr 02 '19 at 13:19
  • 2
    @uhoh you're claiming that a GPS receiver can pick up arbitrarily weak signals. That's impossible. There has to be a threshold below which signals are too weak to be picked up by the receiver. – Hobbes Apr 02 '19 at 21:08
  • 1
    @Hobbes No that's not really what I said. All signals great and small (all direct and all reflected signals from all visible sats, of a given type, e.g. L1) plus all sources of noise are treated as on single analog signal. That one signal is then amplified and AGC'd to an optimum level to avoid saturation, and digitized to a single data stream. Then each Satellite's information in that band is pulled out by a different correlator. For each satellite's signal, the weaker it is the more errors there will be during correlation, until it's so weak that correlation returns only meaningless data. – uhoh Apr 02 '19 at 22:03
  • 2
    @Hobbes what I am saying though, is that if the signals from all visible satellites were much stronger by a factor x, multipath effects for GPS would not be worse, because the AGC would just drop all levels down; all direct and all reflected, by the same given amount x, before digitization. – uhoh Apr 02 '19 at 22:05
6

There is existing technology for this, termed "{Active,Passive} GPS Repeater". It is composed of both an antenna outside and inside, with optionally active or passive components to forward the signal. My introduction to it was within an ocean-going vessel, and it worked fine.

http://www.terrisgps.com/how-do-gps-repeaters-work/

https://www.tri-m.com/index.php/product/tri-m/gps-networking/gps-re-radiators/l1-gps-repeater-kit

https://electronics.stackexchange.com/a/7798/216525

Tyson Hilmer
  • 169
  • 2
  • 12
    Note that a GPS repeater doesn't actually give you indoor navigation - everything receiving the repeated signal computes approximately the same position, that of the outside antenna. (Source: we use these repeaters to be able to do indoor testing of GPS-based equipment that we make.) – jasonharper Apr 02 '19 at 13:19
  • 1
    A cite from your first link: "It should be noted that all the repeaters in a single network will transmit the coordinates of the outdoor antenna and not the position of the repeater unit itself." – Uwe Apr 02 '19 at 13:22
  • @jasonharper If they have chosen to transmit the computed GPS position directly to the GPS receivers (so that crew and cruise passenger easily can see on their phones where they are at sea when under deck) instead of via e.g. WiFi, that speaks for the advantage of GPS, as a dominating standard, over indoor positioning systems requiring specific hardware and software. – LocalFluff Apr 02 '19 at 13:31
4

In general, if you need a local indoor navigation system, then you are better off implementing your own. Generally speaking these are called Indoor Positioning Systems. There isn't yet a standard, but there has been some work to making one. It would be MUCH cheaper, more accurate, and overall just better.

GPS satellites are expensive, they are all space rated atomic clocks, some of the most accurate clocks ever created, and the most accurate in orbit. Making thousands of them would be very expensive, and not really gain a lot.

All that being said, there are a few things that can be done. The way indoor GPS typically works is by estimating your location from other means, and looking for the very low signal strength signal coming at the right time.

PearsonArtPhoto
  • 121,132
  • 22
  • 347
  • 614
  • Low altitude GPS satellites do have conceptual problems with time in sight and Earth's shadow. But I'm pretty sure that putting nuclear reactors on the GPS satellites to up their power (if that would be a solution) is cheaper than installing and maintaining unstandardized indoor systems in billions of buildings. Just changing the beacons' batteries or cabling them to the power grid is a pain. GPS isn't more expensive than EU, Russia, China, Japan and India all have gotten their own (redundant) positioning satellite constellations. – LocalFluff Apr 02 '19 at 13:21
  • @LocalFluff Nuclear reactors in LEO ?!! – Antzi Apr 02 '19 at 15:43
  • @Anzi Nuclear rather as the alternative to having GPS satellites in LEO. With enough power generation they could beam their radio the more powerful from 30 times further away, – LocalFluff Apr 02 '19 at 17:17
  • They may not be cheaper though. Suppose you can either include an extra stage in your mobile phone's chipset to (somehow) pick up GPS indoors, or install an indoor positioning system. The former is cheap (add to a handful of chip designs => solved for most devices within 2 years, similar to adding GALILEO or GLONASS, low cost). The latter needs every building to buy new hardware and install it, effectively billions of individual boxes to buy, install, and run on mains power for the rest of time. – Stilez Apr 02 '19 at 21:06
  • 2
    No satellite from far far away will possibly give the precision that a local system would require. Also, indoor will always suffer from multipath, which will degrade the accuracy further. One of the interesting ideas I've heard in this regard is to use the local wifi signals as a kind of GPS signal, which would be totally doable. – PearsonArtPhoto Apr 02 '19 at 21:16
  • @PearsonArtPhoto I was under the impression that this already existed, with Google trying to determine the exact position of each WiFi router on Earth by measuring their distance differences from GPS-located cellphones? Google Maps can locate my laptop with stunning accuracy even though the laptop's never seen GPS. I'm actually on the other side of my home WiFi router than Google thinks, but the distance from it seems just about right. And it's not just pointing at my home address - which it might hypothetically snipe from my ISP - those two markers are about 10 meters apart. – John Dvorak Apr 03 '19 at 09:38
  • In a shopping mall with each little vendor beams their own WiFi across the whole building, there should be plenty enough data to determine the receiver position once you figure out where each router is -- and there are plenty of people walking the adjoining streets where they can still see the networks.Heck, in the near future it might even be able to tell you in which pocket you have your cellphone in if there's a camera pointing at you. – John Dvorak Apr 03 '19 at 09:45
  • @JohnDvorak A problem with that is that people walking around, and the amount of goods on the shelf, interfere with the strength of the radio signals. Since the WiFi/Bluetooth frequency is the same used by microwaves to cook water. And we and alot of stuff consist of alot of water, the signal strength depends not only on where I am, but on where other things are at the moment. Not that GPS solves that, other than coming from the Heavens unobscured. – LocalFluff Apr 06 '19 at 22:26
2

1. Assuming that your application can (or does) use an external GPS antenna, keep in mind that they are not all the same. They are available in different gains.

I use them for two lightningmaps.org receivers. One is far more sensitive than the other, and I might be able to get away from using it indoors (though I have never tried to).

2. Many have had success by placing GPS antennas next to windows. Glass does not attenuate the signals like foil-backed building insulation, wiring, etc. does.

A GPS receiver connected to an external (outdoor) antenna would determine the position of that outdoor antenna.

Mike Waters
  • 125
  • 6
  • 1
    Question did not ask for GPS indoor receiption only, it was about GPS indoor navigation too. A GPS receiver connected to an external (outdoor) antenna would determine the position of that outdoor antenna. – Uwe Apr 03 '19 at 20:17
  • @Uwe Thanks. Comment edited to include that. – Mike Waters Apr 03 '19 at 20:27
1

You have a couple of big problems here:

1) GPS simply doesn't have the accuracy you need. In most cases the best it could hope for is to figure out what room you're in and that wouldn't be 100% accurate.

2) The more stuff in the way the more reflections become an issue and the more inaccurate your fix is even if you manage to get one. There are special (expensive) high accuracy GPS antennas--but they are utterly intolerant of obstructions.

Note that there is an obvious answer that fails--you can't have an antenna to relay the signals. A GPS receiver does not actually figure out it's location, but rather the location of the antenna used. In a normal civilian use the difference is so small it doesn't matter but when you have a GPS built into something large (say, a ship) the antenna can be a substantial distance from the electronics.

Loren Pechtel
  • 11,529
  • 26
  • 47