6 ms·

Sort of related question: is it possible to get more precise global positioning from a lot more satellites? I was wondering if a swarm of satellites is capable

Sort of related question: is it possible to get more precise global positioning from a lot more satellites? I was wondering if a swarm of satellites is capable of GPS.

jvanderbot 3y ago
I can only answer from theory, but the geometric dillution of precision decreases with N, the number of independent measurements of satellite signals. In practice, there's a floor to the noise which is a reasonable lower bound on your uncertainty. There's also factors that can't really be overcome from more sats, like multipath effects.
jcrawfordor 3y ago
More satellites is mostly better. Multi-constellation receivers for example can obtain higher precision by using some combination of different GNSS. There is a minimum error you can practically achieve this way that has a fair amount to do with the wavelength, but it does help. In most cases, though, faster/more reliable fix will probably be a bigger advantage than precision. When you have a partial sky view, more satellites means a better chance of enough satellites falling in that window for a fix.
- zettabomb 3y ago
  I'm curious, hoe do multi-constellation receivers use them all at once? Are the signals just all the same thing? On some level I know they're just atomic clocks, and the governing equations should work if all of them are synchronized, but I can't help but think that feeding data different constellations in all at once wouldn't produce a good result. Or perhaps you can answer somewhat the opposite question - if the data IS all the same, then why are there multiple constellations and standards? Apart from Beidou, which I believe can actually serve as a radio repeater of some kind, aren't they all doin the same thing?
  - jcrawfordor 3y ago
    The calculation of a fix from GNSS is actually rather complex and the potential complexity is almost unlimited. The most accurate GNSS fixes are obtained from offline solvers that analyze long-duration (e.g. 24 hour) recordings of the GNSS data streams and use computer models to determine the most likely solution accounting for many types of error. This is all to say that there isn't really one approach to calculating a fix from GNSS, and receivers embedded in devices like phones use pretty simplified methods so that they can run in real-time and with a short time to first fix. Most GNSS are largely similar, though, and there are certainly enough papers describing how to perform fixing across satellites of independent constellations so I'm sure it's been implemented in at least higher-end devices. Still, I suspect inexpensive consumer receivers are probably calculating a fix per constellation and then combining those. Fixing across constellations requires combining ephemera, time base, etc. data that varies between them, but all of these are known with high precision, you just have to pull them together. Since most GNSS are mostly the same, new GNSS constellations are usually launched either to improve coverage in a region, or these days with several global constellations, more often for sovereignty reasons. If you have weapons systems that rely on GNSS, it's better to have your own constellation. GPS for example no longer has a wartime selective availability capability (that would deny use by anyone other than US forces) but it did in the past and, if WW3 broke out, it would probably be implemented again. For much this reason different GNSS constellations tend to operate in different bands (not always, though) so that a country can selectively jam adversarial GNSS systems without affecting their own. GNSS has its origins in the military and is still largely a military capability today, with Galileo being the odd one out that focuses more on commercial applications (Galileo's selective availability capability is used to charge a subscription for higher-precision uses). And some GNSS do have "accessory" capabilities, probably the most prominent being Galileo's two-way search and rescue messaging capability that will compliment COSPAS-SARSAT. But that's always been the way with satellites. For example, the GPS constellation carries detonation sensing equipment for nuclear counter-proliferation. It just saves money to combine payloads. One of the reasons to not put GNSS constellations in separate bands is that it simplifies the construction of multi-constellation receivers. Of course, it also makes it more difficult to selectively jam. Some combination of these interests mean that GPS, GLONASS, and Galileo are very close siblings in the radio spectrum. It gets surprisingly political.
SCM-Enthusiast 3y ago
Guessing here but i bet this is one of the line items in the starshield project with the feds, but not a very big commercial concern as GPS is "Good enough" for most applications. However, this would require different hardware on both the sats and the ground stations. You'd prolly need much better coordination algo's too.
LeoPanthera 3y ago
In theory, yes. But it's diminishing returns, and requires more and more capable hardware as you include more. Under ideal conditions GPS is already accurate to 30cm, which is frankly astonishing on a planet of 200 million square miles.
- popol12 3y ago
  cm and square miles in the same sentence You're half-way there, keep it up !
mandevil 3y ago
Define exactly what you mean by "a swarm of satellites"? Each satellite in a GPS style constellation (and I haven't looked into the details, but I believe that all of the major GPS competitors use the same technique- Galileo, GLONASS, Beidou) has its own atomic clock in space and really can only do satellite navigation stuff[1]. Generally speaking, in the medium altitude orbits preferred for this mission, 24 is minimum for world-wide network. Having some more of these specific, large satellites is good: it helps you have more of them in sight at any given moment anywhere in the world, which should give you better accuracy, though having them too close to each other doesn't really help that much, and it helps to have some in reserve so if there is a problem with one you can activate another and quickly fill the gap. According to wiki there are 31 GPS sats operational (plus four in orbit but on reserve), 31 (plus five in orbit but on reserve) Beidou, 23 operational Galileo, and 24 GLONASS operational. So you can see that all of the major operations are at about the same number of satellites for a traditional global satellite navigation system. It might be possible to do a different technique and use a swarm of internet satellites: something like how the TRANSIT (aka NAVSAT) system could, in theory, provide navigation from satellite internet machines. In practice you'd need to know each member of the swarm's orbits very precisely to have sufficient accuracy with their Doppler shift, which means that you prefer station-keeping burns to be infrequent[2], and because they are in low orbit- lots of drag, so they need to regularly burn- that's harder to do (this is why the existing systems use medium orbits). Also, a TRANSIT style operation can't really give you turn-by-turn directions on your cellphone, it was about updating a very powerful submarine inertial guidance system with big honking mechanical gyroscopes every few days, the INS on your phone isn't nearly good enough to handle turn-by-turn directions with TRANSIT update rates. 1: Japan has their own Just Japan system, QZSS, which I have vague memories is different, but I don't remember enough to explain it. 2: The way these systems work, you can never really have more accuracy than you know the satellites location in space. In general, you want ground based systems to track and update orbital parameters after every satellite burn to get the necessary accuracy.
- Dalewyn 3y ago
  >1: Japan has their own Just Japan system, QZSS, which I have vague memories is different, but I don't remember enough to explain it. Japan's QZSS is complementary to America's GPS and also covers Australia and everywhere in-between due to its orbit. For generally practical purposes, it can be considered part of GPS. Any GPS receiver in range can receive its signals.
wkat4242 3y ago
We already have that swarm. Considering most chipsets can receive 4 or 5 GNSS networks with up to 3 bands. That's a lot of signals.
myself248 3y ago
More precise, perhaps. There are a whole lot of GNSS satellites up there already, but they comprise four separate systems, so you need to know a lot of inter-system timing bias information in order to fuse them into one big effective constellation. If you meant a single LEO swarm operated by a single entity, yes that's possible too, and there are proposals to do it, because LEO birds can provide much stronger signals, you'd have a decent shot of it working indoors and stuff. But, the ground-segment necessary to keep track of the clocks on a large number of satellites would be formidable. They'd also be out of view of ground stations for a larger portion of their orbits (since they're so much lower), which means some new work in tracking the clock drift, which is one of the main limitations on accuracy.
grecy 3y ago
Elon tweeted at one point that the swarm of Starlink sats could indeed be used for a GPS-like service, but that it's a low priority and won't happen for a while.

Sort of related question: is it possible to get more precise global positioning from a lot more satellites? I was wondering if a swarm of satellites is capable of GPS.

jvanderbot 3y ago

I can only answer from theory, but the geometric dillution of precision decreases with N, the number of independent measurements of satellite signals. In practice, there's a floor to the noise which is a reasonable lower bound on your uncertainty. There's also factors that can't really be overcome from more sats, like multipath effects.

jcrawfordor 3y ago

More satellites is mostly better. Multi-constellation receivers for example can obtain higher precision by using some combination of different GNSS. There is a minimum error you can practically achieve this way that has a fair amount to do with the wavelength, but it does help. In most cases, though, faster/more reliable fix will probably be a bigger advantage than precision. When you have a partial sky view, more satellites means a better chance of enough satellites falling in that window for a fix.

zettabomb 3y ago
I'm curious, hoe do multi-constellation receivers use them all at once? Are the signals just all the same thing? On some level I know they're just atomic clocks, and the governing equations should work if all of them are synchronized, but I can't help but think that feeding data different constellations in all at once wouldn't produce a good result. Or perhaps you can answer somewhat the opposite question - if the data IS all the same, then why are there multiple constellations and standards? Apart from Beidou, which I believe can actually serve as a radio repeater of some kind, aren't they all doin the same thing?
- jcrawfordor 3y ago
  The calculation of a fix from GNSS is actually rather complex and the potential complexity is almost unlimited. The most accurate GNSS fixes are obtained from offline solvers that analyze long-duration (e.g. 24 hour) recordings of the GNSS data streams and use computer models to determine the most likely solution accounting for many types of error. This is all to say that there isn't really one approach to calculating a fix from GNSS, and receivers embedded in devices like phones use pretty simplified methods so that they can run in real-time and with a short time to first fix. Most GNSS are largely similar, though, and there are certainly enough papers describing how to perform fixing across satellites of independent constellations so I'm sure it's been implemented in at least higher-end devices. Still, I suspect inexpensive consumer receivers are probably calculating a fix per constellation and then combining those. Fixing across constellations requires combining ephemera, time base, etc. data that varies between them, but all of these are known with high precision, you just have to pull them together. Since most GNSS are mostly the same, new GNSS constellations are usually launched either to improve coverage in a region, or these days with several global constellations, more often for sovereignty reasons. If you have weapons systems that rely on GNSS, it's better to have your own constellation. GPS for example no longer has a wartime selective availability capability (that would deny use by anyone other than US forces) but it did in the past and, if WW3 broke out, it would probably be implemented again. For much this reason different GNSS constellations tend to operate in different bands (not always, though) so that a country can selectively jam adversarial GNSS systems without affecting their own. GNSS has its origins in the military and is still largely a military capability today, with Galileo being the odd one out that focuses more on commercial applications (Galileo's selective availability capability is used to charge a subscription for higher-precision uses). And some GNSS do have "accessory" capabilities, probably the most prominent being Galileo's two-way search and rescue messaging capability that will compliment COSPAS-SARSAT. But that's always been the way with satellites. For example, the GPS constellation carries detonation sensing equipment for nuclear counter-proliferation. It just saves money to combine payloads. One of the reasons to not put GNSS constellations in separate bands is that it simplifies the construction of multi-constellation receivers. Of course, it also makes it more difficult to selectively jam. Some combination of these interests mean that GPS, GLONASS, and Galileo are very close siblings in the radio spectrum. It gets surprisingly political.

SCM-Enthusiast 3y ago

Guessing here but i bet this is one of the line items in the starshield project with the feds, but not a very big commercial concern as GPS is "Good enough" for most applications. However, this would require different hardware on both the sats and the ground stations. You'd prolly need much better coordination algo's too.

LeoPanthera 3y ago

In theory, yes. But it's diminishing returns, and requires more and more capable hardware as you include more. Under ideal conditions GPS is already accurate to 30cm, which is frankly astonishing on a planet of 200 million square miles.

popol12 3y ago
cm and square miles in the same sentence You're half-way there, keep it up !

mandevil 3y ago

Define exactly what you mean by "a swarm of satellites"? Each satellite in a GPS style constellation (and I haven't looked into the details, but I believe that all of the major GPS competitors use the same technique- Galileo, GLONASS, Beidou) has its own atomic clock in space and really can only do satellite navigation stuff[1]. Generally speaking, in the medium altitude orbits preferred for this mission, 24 is minimum for world-wide network. Having some more of these specific, large satellites is good: it helps you have more of them in sight at any given moment anywhere in the world, which should give you better accuracy, though having them too close to each other doesn't really help that much, and it helps to have some in reserve so if there is a problem with one you can activate another and quickly fill the gap. According to wiki there are 31 GPS sats operational (plus four in orbit but on reserve), 31 (plus five in orbit but on reserve) Beidou, 23 operational Galileo, and 24 GLONASS operational. So you can see that all of the major operations are at about the same number of satellites for a traditional global satellite navigation system. It might be possible to do a different technique and use a swarm of internet satellites: something like how the TRANSIT (aka NAVSAT) system could, in theory, provide navigation from satellite internet machines. In practice you'd need to know each member of the swarm's orbits very precisely to have sufficient accuracy with their Doppler shift, which means that you prefer station-keeping burns to be infrequent[2], and because they are in low orbit- lots of drag, so they need to regularly burn- that's harder to do (this is why the existing systems use medium orbits). Also, a TRANSIT style operation can't really give you turn-by-turn directions on your cellphone, it was about updating a very powerful submarine inertial guidance system with big honking mechanical gyroscopes every few days, the INS on your phone isn't nearly good enough to handle turn-by-turn directions with TRANSIT update rates. 1: Japan has their own Just Japan system, QZSS, which I have vague memories is different, but I don't remember enough to explain it. 2: The way these systems work, you can never really have more accuracy than you know the satellites location in space. In general, you want ground based systems to track and update orbital parameters after every satellite burn to get the necessary accuracy.

Dalewyn 3y ago
>1: Japan has their own Just Japan system, QZSS, which I have vague memories is different, but I don't remember enough to explain it. Japan's QZSS is complementary to America's GPS and also covers Australia and everywhere in-between due to its orbit. For generally practical purposes, it can be considered part of GPS. Any GPS receiver in range can receive its signals.

wkat4242 3y ago

We already have that swarm. Considering most chipsets can receive 4 or 5 GNSS networks with up to 3 bands. That's a lot of signals.

myself248 3y ago

More precise, perhaps. There are a whole lot of GNSS satellites up there already, but they comprise four separate systems, so you need to know a lot of inter-system timing bias information in order to fuse them into one big effective constellation. If you meant a single LEO swarm operated by a single entity, yes that's possible too, and there are proposals to do it, because LEO birds can provide much stronger signals, you'd have a decent shot of it working indoors and stuff. But, the ground-segment necessary to keep track of the clocks on a large number of satellites would be formidable. They'd also be out of view of ground stations for a larger portion of their orbits (since they're so much lower), which means some new work in tracking the clock drift, which is one of the main limitations on accuracy.

grecy 3y ago

Elon tweeted at one point that the swarm of Starlink sats could indeed be used for a GPS-like service, but that it's a low priority and won't happen for a while.