Shop OBEX P1 Docs P2 Docs Learn Events
How does GPS determine elevation? — Parallax Forums

How does GPS determine elevation?

lardomlardom Posts: 1,659
edited 2010-09-13 23:11 in General Discussion
This is just to satisfy my curiosity. While stuck in traffic yesterday my thoughts drifted back to something that's puzzled me for a while: I kept wondering how the minimum three satellites could find a GPS location because you could have a large triangular area where three circles intersect. Geometry doesn't answer the question. Then the light goes on in my head...The GPS reciever is located on the circumferences of those three circles. (Traffic jams are a good thing.)
This is what I understand so far: The satellites have synchronised atomic clocks. The GPS reciever has a quartz clock that is synchronised to the atomic clocks. The reciever measures the lag time when the satellite transmits the current time and its location.
OK, from circles to spheres and from a spot on the map to elevation on the planet. How does the reciever do this?

Comments

  • bill190bill190 Posts: 769
    edited 2010-09-12 08:20
    How about using 4 satellites...
    http://edu-observatory.org/gps/height.html
  • HollyMinkowskiHollyMinkowski Posts: 1,398
    edited 2010-09-12 11:01
    Well, you get a time signal and a precise satellite location
    from each of three or more satellites. Using your precise time
    on the ground and the data from the sats your gps calculates
    your exact location in 3 dimensions since it knows how far each sat
    was from you when the readings were made. If you know your exact
    distance from 3 objects then you can calculate where you are
    in 3 dimensional space. (There is a special mathematical case where 3 objects
    cannot be used to determine your location in space but this does not happen
    with gps)

    Special Relativity must be taken into account. The clocks on the satellites
    slow down due to their speed. The clocks on the ground slow down even
    more due to the Earths gravity well. The end result is that the clocks in orbit
    must be slowed down just a tiny bit more to match the gps clocks on the
    ground. The processor in your gps device takes all these things into account
    when it makes its calculations.
  • ercoerco Posts: 20,257
    edited 2010-09-12 16:21
    Nice description, Holly! Speaking of slow clocks, I have a cheap China Ebay "Hello Kitty" clock that runs a little slow. Per your description, can I reverse the polarity and make it into a GPS unit? How would I do that? Can I use a relay and a BASIC Stamp?
  • Martin_HMartin_H Posts: 4,051
    edited 2010-09-12 18:12
    I understand the location of an object in three space. But altitude is your height above sea level which has always bothered me. How do they define what sea level should be? Do they use some kind of idealized ellipsoid or do they have a better idea of where sea level should be?

    The reason I ask is that sea level isn't really level as it varies based upon the density of the rock that underlies the ocean.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2010-09-12 19:40
    An approximation to the geoid is used, depending on the chosen geodetic system (e.g. WGS 84).

    -Phil
  • HollyMinkowskiHollyMinkowski Posts: 1,398
    edited 2010-09-12 23:27
    erco, you are so funny :-)
    Your daughters sure have a fun dad to play with :-)
  • lardomlardom Posts: 1,659
    edited 2010-09-13 10:29
    Phil, You gotta admit a geoid sounds like something a surgeon should remove. :) It looks to me like elevation is determined from a standardized model after the reciever has been pinpointed. I know you can answer this question; Is a Propeller as capable as the processor in a GPS?

    I attached a graphic that shows what 'clicked' in my head.
    360 x 563 - 33K
  • David BDavid B Posts: 592
    edited 2010-09-13 11:16
    I know it's harder to visualize, but I think you need to imagine intersecting spheres rather than circles. The spheres represent the distance of the receiver from each satellite.

    Two three-D spheres intersect in a two-D circle; the third sphere intersects that circle at two points, and usually by knowing that your location is near the surface of the Earth gives enough info to eliminate one of those two points, yielding your GPS location.

    A fourth satellite is usually needed to be able to mathematically deduce the receiver clock time offset as well as the location from the collection of data values, allowing GPS receivers to work with just a relatively cheap clock.
  • localrogerlocalroger Posts: 3,452
    edited 2010-09-13 12:49
    Lardom, let me try to put it another way. Let's say you have three points on a tabletop. These are your satellites. And at each point, you have a hinged rod that is fixed to the point but free to rotate in any direction in three-dimensional space. Those are the distances returned by the GPS receiver telling its distance from each satellite.

    Now, let's just consider two of those. Take two of those rods and fasten them together at the far end -- the receiver end. Now that composite point can only swing in a circle, because it's constrained by two satellite fixes. Your receiver must be somewhere on that circle.

    Now, hook up the third rod. There are only two places you can do that -- one point above the table, and one underneath it if you can swing the rods down there. Your receiver knows it's at one of those point locations, and since one of them is probably near the surface of the Earth and the other far out in space it usually guesses lucky.

    The GPS receiver does need to know the shape of the Earth too, since it's not exactly spherical, but there is a standard geoid model used for that. It's not exact though which is why altitude information isn't considered as accurate for GPS as position.
  • lardomlardom Posts: 1,659
    edited 2010-09-13 23:11
    Thanks, everyone. Understanding GPS recievers was something I really wanted to know. I bought my first gps, I guess, about three years ago having to find my way around a rural area in Virginia. There was no map book I could buy, there was only the hotel phone or Mapquest. I still got lost. I reluctantly spent $300 and put in my home address. The device said "In forty yards turn right..." I drove around with a grin on my face for two days straight.
Sign In or Register to comment.