Europe's GPS Rival [Archive] - Thoughts Media Forums

Rocco Augusto

07-17-2006, 08:40 PM

<div class='os_post_top_link'><a href='http://news.com.com/2300-1033_3-6011169-1.html' target='_blank'>http://news.com.com/2300-1033_3-6011169-1.html</a><br /><br /></div>"<i>The European Union launched its first Galileo navigation satellite, the Giove-A, on Dec. 28 in a program to end Europe's reliance on the United States' Global Positioning System, or GPS. The European system, which is expected to go into service in 2008, aims to have an accuracy of 3 feet or less--five times better than the current GPS system... The fully deployed Galileo system planned will consist of 30 satellites, positioned in three circular Medium Earth Orbit planes.</i>"<br /><br /><img src="http://www.smartphonethoughts.com/images/augusto-2006.07.16-gps.rival.jpg" alt="User submitted image" title="User submitted image"/><br /><br />Let the battle of the satellites begin! If there is one thing I always was amused of with GPS is just how wrong it could be at times, like driving down your favorite road but your GPS screen says you're driving in a river. Being five times better than current systems in place it is going to be interesting to see how this all pans out. We might even have a new acronym on our hands here. ;)

Sven Johannsen

07-18-2006, 05:05 AM

I apologize in advance, but I feel strangly compelled to say something about accuracy and precision. Accuracy being how right it is and precision being how granular a measurement you have. They are different. For example, I am 6ft tall. Most of you assume I mean 6' 0", but I didn't say that. I could be anything from a bit over 5.5' to just under 6.5' and 6' is still accurate, just not very precise, I have a 1 ft potential variance. I could say I am 7'3.5" tall, which is very precise, varying only in 1 tenth of one inch, but it is, in this case, not very accurate.

What has this to do with GPS? Well how does your PPC/GPS show your location? Pocket Streets on my PDA-N says I am at XX.9483N and XXX.7501W. So my variance could be .0001 degree. So how far is that? Well the earth is about 24000 miles around and 360 degrees. So 1 deg is about 66.67 miles (24000/360), or 352000 ft (66.67mi*5280'). So .0001 times 352000 is 35.2 ft. So when my GPS says I am at xx.9483N, I could really be anywhere between xx.9482 and a 1/2, and xx.9483 and a 1/2. So I have a .0001 degree or 35' slop just because of my precision. What I'm saying is that if my display shows xx.9483, I could walk 35 ft before the display changes to xx.9484, and that would be right. I would have to show xx.94830 before walking 3.5 ft would change the display to xx.94831.

What happens if we do Deg.min.sec. Well 1 deg is still 66.67 miles. Divide by 60 (min in deg) = 1.111 miles. Divide by 60 again (sec in min) and a second is .0185 miles, times 5280 = 98' about. So degree,min,sec get you within 98 ft. Most such displays actually get you to sec.x or 104,32,45.5 which gets you down to about 10 ft, but you still have a 10 ft slop just because of the precision. If you were precise to hundredths of a second, instead of a tenth, you should be down to a foot.

Can we do that, yea, technically the GPS system can, but there are numerous things that afect how accurate things are. You can google how it all works, but basically you need at least three satellites to get a 2D fix, four for 3D. You do some of the math based on how long it takes the signal to get from the satellite to your GPS, but that requires that you know when it sent the signal and when you got it. The satellites have a very accurate clock, but your GPS doesn't. So the delta times are actually part of the calculation. Suffice it to say, you are working 4 to 5 variables in as many equations. If you have taken linear algebra, you know this is a bear to do, but it is made easier by the brute force of a computer. It still takes time though. It takes more time if you want the answer down to .00001 than if you are satisfied with .0001. There are natural effects that work against you getting good numbers to start the calculations too. At some point you have to decide if it is worth doing calculations to 5 decimal places starting with values you are only sure of to 4 decimal places.

Can you get more accuracy than we see with our units. Sure you can. http://www.eclipse.net/~essco/gps/promark.htm, but look at the price. To a point it's just a matter of money. How much do you want to pay for the processor that can do it, and how big are you willing for it to be?

Maybe the new European system has a different process that makes allowances for some of the issues that affect the current GPS system. it is after all decades newer. They are planning to sell subscriptions though, so there better be some reason to pay for it.

You know Russia (Soviet Union) has had their nav sat system, Glonass, up for a while too.

Sorry about this, but I had nothing better to do ;)

P.S. Yes the math is a bit different north/south than east/west, depending where you are, but the basic concepts are there.