GPS better than GM

Posted by Jim Pruett on Sep 26 2007 under   

I got another speeding ticket. Better me than you, right. This one got me thinking:

Which is more accurate, GPS or a car speedometer? What do you think?

There is plenty of σ-math available relating to GPS. My guess is no one knows the real truth about GPS accuracy, so as my old friend Dr. Doug Wiley said, everyone just bullsh-ts with Mathematics. (Except for that one man from Lincoln Labs that invented it!)

To prove my point, heres what the last grad student said on Wikipedia.

"The position accuracy is primarily dependent on the satellite position and signal delay... Position accuracy can be improved by using the higher-chiprate P(Y) signal. Assuming the same 1% bit time accuracy, the high frequency P(Y) signal results in an accuracy of about 30 centimeters."

Ok, so that needs some work. Forget that for a moment. Heres the explanation I will propose. Using a picture I butchered from www.nujournal.net, you can obviously see that speed is very wiggly.

If I take two measurements, I get a wiggly answer.

The other part is hills. Do hills affect GPS calculations? One reference used skydiving to determine this. I think his proof went to 11. Another reference proved that the vertical error is twice as bad as the horizontal error. That made sense to me, we drive in the same plane, parallel to the satellites, unless its Friday night where we move in a monotonically wiggly fashion.

The bottom line for me is that using my Delorme software, the lat has 4 digits after the decimal and the height has 2 digits after the decimal, so the accuracy is 0.01mph. Q.E.D. indeed.

This fits the consensus I found for Vanilla GPS, which is much better than your tire based +-2mph speedometer. So if you want precision driving, get a GPS and embrace the wiggly factor up a hill for sure.

Can anyone help me slow down?

-Jim

 

Rate this article:

Rating: 0 | Votes: 0

posted by Jim Pruett
Jim Pruett has designed consumer electronics that all have one common theme, they are patentable. Every processor from DSPs to the PIC to the Starcore. Every language from Python to uClinux. He started out where every EE starts out, at Motorola, then gradually through failures including readynote.com, he emerged as an inventor. After a six year stink in LA, Jim ended up in Memphis, TN with his wife and toys.

Previous post by Jim Pruett: PIC rules, ARM drools
Next post by Jim Pruett: Today is national Blog about the environment day.
all articles by Jim Pruett

I would say you are comparing oranges and orange juice.
Your car has some transducer that translates the transmission speed to miles per hour. Mechanical devices wear in a known fashon and are reasonable predictable.

There are several factors that affect the GPS idea of position as it is calculated each second by most commercial GPS devices. The speed display, if going to be the difference in position over two measurements (in miles) times 3600 to get miles per hour. A position error of 5m leads to an instantaneous error of 4 mph or so. Your GPS device may filter the displayed speed value from several measurments before putting it on the screen. I believe most do this by asking if you are using the device on a car, a plane or a boat. This means that you do not have your exact (radar) speed, just an average over some periiod of time.

Hills can have an effect on GPS position accuracy. If you are close to a large hills, you may get a signal reflecting off of one hill while hidden from LOS by another. This would introduce a time delay in the signal arrival that could cause a larger error than nominal. Also, your GPS may be set to only find a 2D solution, and changing altitude increases error in that solution space. Tall buildings also tend to block the signal and give strange RF reflections.

GPS is not magic, but it is a lot better than using a sextant.