The Unofficial BMW M5 Messageboard (m5board.com) - View Single Post

tonyo · 20th December 2001, 00:38

The accuracy of civilian GPS was enhanced a couple years ago when the selective availability feature was turned off, thus giving us all the accuracy that the civilian channels of the GPS can provide. The accuracy of the information transmitted by the GPS satelites is probably much better than +-10 ft, but atmospheric distortions and other factors will impair that.

As far as altitude is concerned, there are really two major issues.

The first problem is that the GPS system doesn't know or care where sea level is. The GPS receiver independently uses a mathematical formula to estimate sea level at any given position on the planet. The problem is that the planet (which is irregular and not spherical) doesn't bother itself to precisely conform to any known mathematical formula. That difference is a source of error, since your "altitude" is a distance calculation from your 3D position to that imaginary sea level.

The second problem is harder to explain.

In principle, the GPS system is perfectly capable of good 3D positioning, but because it works by distance triangulation, to get the best 3D positioning, you want GPS signals coming from all directions, or more precisely, you want a GPS signal to be substantially aligned with every direction that you care about.

However, since the GPS satellites are in relatively low orbits, there are few that are nearly directly overhead at any given point in time. Further, all the GPS satellites very near to or below the horizon are unable to help. This results in a sort of "saggy trampoline effect" where positions along the surface are well known, but elevations are not.

Imagine that you had well-calibrated strings running to a few "stationary" satellites, all somewhat low in the sky. The spot where you are holding all of the strings in one hand is your position, which you can calculate by reading the markings on your strings and doing some math. Any movement along the ground causes changes in at least one of the strings, which gives you great "2D" position information. However, if none of your strings is going "up", you can't measure small changes in altitude very well, since you are just moving perpendicular to all of the strings and the numbers don't change. But when you do have a satellite nearly overhead, your altitude can be more precisely determined, subject to the first error described above.

20th December 2001, 00:38	#5 (permalink)
tonyo Addicted Member (>300 posts) Join Date: Jan 2001 Location: Austin, TX, US Posts: 389 Thanks: 1 Thanked 0 Times in 0 Posts	The accuracy of civilian GPS was enhanced a couple years ago when the selective availability feature was turned off, thus giving us all the accuracy that the civilian channels of the GPS can provide. The accuracy of the information transmitted by the GPS satelites is probably much better than +-10 ft, but atmospheric distortions and other factors will impair that. As far as altitude is concerned, there are really two major issues. The first problem is that the GPS system doesn't know or care where sea level is. The GPS receiver independently uses a mathematical formula to estimate sea level at any given position on the planet. The problem is that the planet (which is irregular and not spherical) doesn't bother itself to precisely conform to any known mathematical formula. That difference is a source of error, since your "altitude" is a distance calculation from your 3D position to that imaginary sea level. The second problem is harder to explain. In principle, the GPS system is perfectly capable of good 3D positioning, but because it works by distance triangulation, to get the best 3D positioning, you want GPS signals coming from all directions, or more precisely, you want a GPS signal to be substantially aligned with every direction that you care about. However, since the GPS satellites are in relatively low orbits, there are few that are nearly directly overhead at any given point in time. Further, all the GPS satellites very near to or below the horizon are unable to help. This results in a sort of "saggy trampoline effect" where positions along the surface are well known, but elevations are not. Imagine that you had well-calibrated strings running to a few "stationary" satellites, all somewhat low in the sky. The spot where you are holding all of the strings in one hand is your position, which you can calculate by reading the markings on your strings and doing some math. Any movement along the ground causes changes in at least one of the strings, which gives you great "2D" position information. However, if none of your strings is going "up", you can't measure small changes in altitude very well, since you are just moving perpendicular to all of the strings and the numbers don't change. But when you do have a satellite nearly overhead, your altitude can be more precisely determined, subject to the first error described above. __________________ '08 E60 Space Gray M5 '01 E39 Le Mans Blue M5 '99 550 Maranello '98 Supra Last edited by tonyo; 20th December 2001 at 00:41.