Talk:Error analysis for the Global Positioning System

Source for UERE table
In current version of page, there is a table called "Sources of User Equivalent Range Errors (UERE)" that is presented without reference. What is the source of this table? 198.123.51.133 (talk) 18:45, 16 February 2016 (UTC)

This section seems to contradict the GPS SPS Performance Specification referenced at the bottom of the article. Looking back through the edit history shows that the multiplier of 3 was applied to the estimated sigma values in late 2013. Although the contributor referenced a web page in support of this change, that reference doesn't seem to actually support the idea. Perhaps this was vandalism. It would be good if an expert could revisit this section, amending the error values if appropriate, and adding a reliable source in support of the presented data. 101.165.107.222 (talk) 10:03, 28 June 2016 (UTC)

Split-out
Folks, this article contains content copied from the Global Positioning System article. At the moment the content is mostly identical to part of that article. The intent, of course, is that this content in the main article should be summarized with a link added to this article. However, I don't want to make such a major change to the main article without giving interested parties an opportunity to discuss. --Mcorazao (talk) 02:58, 6 June 2010 (UTC)

Grewal (2001), p. 103
'Before it was turned off on May 1, 2000, typical SA errors were about 50 m (164 ft) horizontally and about 100 m (328 ft) vertically’ (‘Grewal (2001), p. 103’)

Looking at 'Global Positioning Systems, Inertial Navigation and Integration' by Mohinder S. Grewal, Lawrence R. Weill and Angus P. Andrews, 2001, p. 103 says:

'In the GPS SPS mode, the SA errors were specified to degrade navigation solution accuracy to 100m (2D RMS) horizontally and 156m (RMS) vertically.'

I can't find any reference to the 50m horizontal / 100m vertical that is quoted in the article. Does anyone know where these values come from? Is it a different Grewal (2001)? (How many Grewal (2001)s on GPS can there be?) I'm loath to change it without knowing where it came from, but the values just don't seem to be quoted correctly. --82.70.156.254 (talk) 13:09, 11 September 2011 (UTC)


 * I suspect the first source is using a different metric than RMS (maybe a different book edition?). I think you should change it and note the edition.
 * 96.93.212.81 (talk) 17:40, 30 April 2021 (UTC)

Error bounds
Article says that civilian GPS accuracy is typically 15m or with the modern technology and under the clear sky is 5m. But what about in the typical city, when building reflections are in effect. I am sure there must be some data on what errors they may typically introduce. I for example find that at particular locations in the city systematic GPS error is 50-100 meters. Yurivict (talk) 02:03, 30 May 2012 (UTC)
 * There certainly is information about that, but it is difficult to put a "typical" number on it. Cell phone antennas are real shitty, urban canyons (where you not only have multipath but also a higher DOP since you only see satellites in one plane) can be very different, internet connected GPS's that are fed databits can pick up much weaker signals, if the receiver have access to other sensors (inertial for example) it can do a better job, and so on.
 * 96.93.212.81 (talk) 17:45, 30 April 2021 (UTC)

Insert at top of article
An insert at the top of this article says, "This article is written like a personal reflection or essay rather than an encyclopedic description of the subject. Please help improve it by rewriting it in an encyclopedic style. (May 2012)". This is a rather vague comment which appears to have no clear meaning. I think neither the commenter nor anyone else has the vaguest idea what this jumble of words mean. Also the commenter shows his failure to understand the meaning of the phrase encyclopedic style. Encyclopedic doesn't refer to a style, it instead means comprehensive. The commenter, Khazar2, has not offered the slightest hint as to what he is talking about either on this Talk Page, his Talk Page or on his User Page. RHB100 (talk) 20:24, 16 January 2013 (UTC)
 * Much of the text appears to have been lifted from a commercial text book or a technical manual. There may be some copyright issues with some of the text here, when it talks about "chapter 11" and reads like a school text book from that point on, it does lend some possibility that some of the text here was lifted from a commercial publication. SoftwareThing (talk) 23:44, 21 June 2018 (UTC)

Relativity corrections and relative times.
I think it is misleading how right now it is phrased that 38 microseconds a day coming from corrections due to relativity would result in 10 km per day error. This time offset is almost the same for all satellites and it is only their relative times what are used for positioning anyway, so these offsets would largely cancel out. It is said as much in the next sentence, but so indirect, that an average reader is likely, I think, to read this passage as "without relativity adjustments GPS would skrew up by 10km a day (so the adjustments are made by GPS)". It seems to me quite misleading, and I think it can be phrased without casting such impressions. What do people think about this? L3erdnik (talk) 01:14, 21 April 2020 (UTC)


 * See : "In a full day, the accumulated error of 3.8 10-5 s would correspond to a position error of (3.8 10-5 s).c = 11.4 km." I have added the source to the sentence, changed 10 km to 11.4 km, and slightly reworded to match the source: - DVdm (talk) 06:53, 21 April 2020 (UTC)


 * Well, what can I say, I have no reason to doubt the -physics- displayed in this -physics- book, as I said, the arithmetic seems ok there, what I am objecting to, is the presumed (in this -physics- example) the -engineering- design of how GPS works. It is true that the relativistic corrections amount to ~10 km/day change in measured -pseudorange- (as correctly stated in the article), but to work out the position, the system uses not pseudoranges - which, because of the receiver clock bias, are by themselves useless anyway, as indicated by the part "pseudo" - but rather differences of pseudoranges, to eliminate that bias. And by the same token, any bias same for all satellites - like the discussed relativity effects - is eliminated by the design of the calculation, as correctly stated in the next sentence of the article: "This initial pseudorange error is corrected in the process of solving the navigation equations" (not by satellite clock attunement, but just the way the receiver does its math). I don't dispute the usefulness of the clock attunement, but its purpose and effects are to be derived from a text on GPS design, not from one that it has different subject and just uses this effect as a setting for demonstrating some point, seemingly without making sure that not only that point is correct, but the whole example. It is like finding a perfectly fine arithmetic book having an example "How to compute, how many mosquitos weigh as one elephant if an elephant weighs 10 tons and mosquito - 1 gram? We need to divide 10 tons = 10000000 grams by one 1 gram and get 10000000." and use it as a source. As a math problem (what the book deals with) it is totally fine, but since the weight of a mosquito is nowhere near 1 g, it has little to do with reality. L3erdnik (talk) 14:31, 22 April 2020 (UTC)

Error analysis for the Global Positioning System (Calculation of time dilation)
Only the time dilation effects due to relative velocity/speed are discussed here – that is the speed measured or calculated between the two IFRs - which in this case is the speed between satellite transmitter and the receiver on Earth. The 3,874 m/s orbital speed quoted as being relative to the Earth's centre has no bearing on the calculations whatsoever. Note that I put the word quoted in there, as that is not the speed relative to Earth’s centre. If the orbit was perfectly circular, the relative speed would be zero, but because it is slightly elliptical, with an eccentricity of about 1%, the speed varies from plus to minus 38.8 m/s over each orbit. The calculations above should use plus to minus 38.8 m/s instead of the 3,874 m/s used, with the complication that it is a varying figure depending on the phase of the orbit. To further add to the complications, the speed of the receiver should be factored in, which in the case of anything faster than a car, such as a jet aeroplane, can be a factor of 10 or more, and can itself be varying in both magnitude and direction. Tom Hollings — Preceding unsigned comment added by 92.30.110.196 (talk) 15:10, 20 July 2020 (UTC)