[MossyRocks]

The Google Maps app on my Android phone has an option to download data to cover areas with spotty coverage. Only good for about two weeks.

As UncleVanya stated that is the actual map data not position info and not any augmented GPS or GPS almanac updates. I use that feature all the time as I don't actually have a data plan but I also use other mapping apps that offer lots of useful features (Avenza maps with the MN DNR maps, or Navit) as well.

[Michail_P]

Interesting to know how this works. I know that some surveying stations can actually achieve an accuracy of 1.5-2mm. Pretty good if you think about it.

[AstroDave]

was under the impression that GPS worked on triangulation - solving the distance to three known points (satellites) to determine the location in a 3D plane. As the distance is determined based on a time difference, a fourth satellite is required for this.

You basically have four unknowns (x, y, z, t) and therefore need four "equations" so to speak to solve. More satellites provide more accuracy, with increased certainty.

You got it! For a complete position and time solution, you have 4 variables --> need 4 independent measurements (the time-delays wrt the satellites; in GPS-speak: "pseudo-ranges") to solve for the 4 variables !!

If you think your clock is good (enough), you can get (X, Y, Z) (or, equivalently, Lat, Lon, Altitude - the overall GPS system includes some assumptions about the shape of the Earth: the geoid and sea level). Conversely, if you fix/quess your altitude, you can get time/lat/lon.

[Paul the Sunman]

5 m accuracy and 1 min acquisition time is good enough for me, but all these enhanced capabilities are certainly impressive.

[subsea]

I'm confused. Maybe I am grossly misunderstanding the functionality of how GPS works, but I cannot see how a position can be found with a single satellite.

I was under the impression that GPS worked on triangulation - solving the distance to three known points (satellites) to determine the location in a 3D plane. As the distance is determined based on a time difference, a fourth satellite is required for this.

You basically have four unknowns (x, y, z, t) and therefore need four "equations" so to speak to solve. More satellites provide more accuracy, with increased certainty.

The posts above by @AstroDave; and @MossyRocks; are quite interesting to read, I had no clue that accuracy to such fine precision was able to be met these days. And reading the phase of the wave seams pretty clever, when I have some free time I think I'll read more into this.

The early civilian GPS systems made many assumptions regarding things such as starting location and other things which were intended to fill in holes resulting from less than ideal signals (as AstroDave enumerates above). These assumptions, if I remember correctly, enabled the multiple equations to be solved even if all if the necessary data wasn't "known", but would be ignored if the system got signals from enough satellites. Depending on the amount of data obtained from the satellites, the accuracy of the solution would degrade "gracefully" until it reached the point of total nonsense. The military systems were less forgiving in that any piece of missing data resulted in total failure of the system to determine a location. As time progressed, the sensitivity of GPS receivers and the number of GPS satellites available to supply signals, improved to the point where this problem, and the kluge used to address the problem, became pretty much irrelevant.