51
$\begingroup$

This Newsweek article states that "GPS...will likely be nonexistent near the eclipse zone" because of all the people going to that area. While it makes sense that cell towers might be overloaded and it might be impossible to download maps or conduct Internet searches that find coordinates for an input address, and one shouldn't plan to be streaming music/videos over mobile connections during that time, why would GPS be affected?

I thought GPS didn't require communication from devices to satellites, it only required the devices to listen and compute, and thus the system was scalable to a very large number of devices, even when those devices are relatively close together.

$\endgroup$
  • 20
    $\begingroup$ Poor reporting. The source is either misinformed or was talking about E911. $\endgroup$ – pericynthion Aug 6 '17 at 20:59
  • 8
    $\begingroup$ It is true that a GPS device is a pure receiver. The GPS system was designed this way, it should be impossible to determine the position of GPS device by listening to the communication between the device and the satellites, thus there is no such communication. As a consequence, the number of GPS devices per square mile of earth surface is not limited. $\endgroup$ – Uwe Aug 6 '17 at 21:30
  • 1
    $\begingroup$ There might be one issue though: I guess Ionosphere properties are changing rapidly due to missing sun light. Hence, location accuracy might be worse than usual. WAAS might not be able to cover it because it is a local and fast phenomenon. $\endgroup$ – asdfex Aug 7 '17 at 14:16
  • 32
    $\begingroup$ If some people call their browsers "the internet", calling their map apps "GPS" is not in the least bit surprising. $\endgroup$ – wedstrom Aug 7 '17 at 15:23
  • 2
    $\begingroup$ @wedstrom Not all map apps need internet access at use time (e.g. MapFactor Navigator does not). $\endgroup$ – WBT Aug 7 '17 at 17:00
62
$\begingroup$

GPS isn't affected by demand, as it is transmitting only from satellites, and the receivers only receive, they do no transmit to the satellite at all. At best, there is a slight degradation by having antennas in really close proximity. A million man march, each with a GPS device, might cause some degradation as each device will absorb a bit of the energy that others might. If it were to happen, it would only absorb the signals at low angles, which aren't really required, although they do help to get a better solution. There could also be some noise generated, as receivers can transmit a very small amount of signal in the range of the intended signal as generated by their oscillators. But in large part, it won't be degraded, and at worst case, would only be reducing the number of satellites slightly. And even that effect is probably more due to the large number of people around you, and not the electronics which they possess.

Differential GPS, on the other hand, might have some degradation. This requires that GPS corrections be sent out, and the network traffic that these are sent out on might not work as well. Still, that would only lead to uncertainty in the highest order, it would still be able to figure out which street you are on.

From the article comes the following quote:

print out directions since GPS (especially Google Maps) likely won’t be an option

Many people associate GPS with turn by turn directions from something like Google Maps. And that has the potential to be really affected. I suspect the author made the same mistake.

$\endgroup$
  • 4
    $\begingroup$ However, if people download an offline map application like MapFactor Navigator, and download maps for the appropriate states in advance, they should NOT be thinking that the GPS signal will be crowded out. (Paper maps are still a good idea in any case, but not that much more in this case than others.) $\endgroup$ – WBT Aug 6 '17 at 22:05
  • 10
    $\begingroup$ The wavelengths of GPS L1 and L2 are 19cm and 23cm respectively, that's 7.5 and 9.2 inches, which is a heck of a lot closer than people are going to be holding their GPS units even in a crowd. The speculation about mutual interaction and signal loss due to receiver proximity is not based in science. The rest of the answer is great! $\endgroup$ – uhoh Aug 7 '17 at 2:47
  • 3
    $\begingroup$ It doesn't apply, as the GPS devices are only receiving, not transmitting. Does quality degrade based on how many people listen to a radio station? $\endgroup$ – PearsonArtPhoto Aug 7 '17 at 11:19
  • 4
    $\begingroup$ Cold fix probably will be affected. You can't download the ephemeris at the last moment over GSM or WiFi if there are ten thousand people around you trying to do the same. $\endgroup$ – John Dvorak Aug 7 '17 at 16:47
  • 1
    $\begingroup$ @JohnDvorak for large events that aren't deserving of permanent equipment (i.e. Lollapalooza is 4 days/year, not the Cubs' Wrigley Field which is 80-90 days/year), Cells On Wheels (COWs) are set up in advance to handle the overwhelming network traffic. $\endgroup$ – Nick T Aug 7 '17 at 23:10
39
$\begingroup$

As PearsonArtPhoto says, it's not the GPS protocol itself that causes the problem.

Cell phones use Assisted GPS, where cellular data is used to speed up obtaining a GPS fix. This should be just a few kb per session though.

Many mapping applications also download map data as you go along, again causing lots of network traffic.

This may be an aberration, but I'm leaving it here for future reference: in my own tests an iPhone went through 100 Mb of data just for AGPS in 1 hour.

$\endgroup$
  • 4
    $\begingroup$ This is the correct answer. The article specifically puts it in that context: "Cellular service towers aren’t meant to handle the capacity of an additional half-million to a million people per state. Cellphone, GPS and smartphone internet services will likely be nonexistent near the eclipse zone..." +1 $\endgroup$ – rasher Aug 6 '17 at 23:51
  • 5
    $\begingroup$ Thats a lot. Are you sure it is just the agps and not maps (the foot version as even vector maps are not that big usually)? If it were so much, many Pokemon go players would be broke long time ago. $\endgroup$ – jkavalik Aug 7 '17 at 6:09
  • 21
    $\begingroup$ 100 Mb per hour can't be right. The whole almanac has just 15 kBit, plus a few more kB for WAAS/EGNOS data. The alamanc doesn't need regular updates, WAAS should be updated every few minutes. $\endgroup$ – asdfex Aug 7 '17 at 8:26
  • 1
    $\begingroup$ That's crazy. Like asdfex says, a SUPL response isn't more than a few kilobytes, and once the GPS lock is acquired there's no use in making any more requests for the acquisition parameters (so no continuous use there), and the ionospheric stuff doesn't change rapidly enough to need frequent updates. Either whatever you're measuring is the data usage of some other feature, or Apple is being terribly wasteful of their customers' money. $\endgroup$ – hobbs Aug 7 '17 at 16:32
  • 3
    $\begingroup$ 100 Mb per hour is indeed insane. I implemented ephemeris distribution for a well-known Portable Navigation Device vendor, and that was measured in kilobytes per week. (ephemeris prediction works quite well). You have a clock error when it's 100 Mb per hour - the device thinks it always has stale data, even after receiving current data. $\endgroup$ – MSalters Aug 8 '17 at 11:12
2
$\begingroup$

It isn't the GPS, but the applications using this data. On your phone, you generally wouldn't interact with GPS, but with a Map, in some form. This map data is not stored on your phone. The applications using the GPS are what puts the load on the network.

$\endgroup$
  • 3
    $\begingroup$ "This map data is not stored on your phone." Why would you think so? 3 out of 4 mapping apps on my phone do have the data on the phone. The remaining one has the option cache the maps too. $\endgroup$ – Vladimir F Aug 8 '17 at 20:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.