Doug, I think the 3 meter thing was a few years ago, because the government intentionally screwed up the accuracy of the signals the satellites sent to civilian gps for security reasons. I heard they made it a lot more accurate a few years back though, so I think the 3 meter number now would only come into play, as Tim said, when you're talking about really poor atmospheric conditions. I think with good conditions, GPS today is supposed to be pretty damn precise. I've heard accuracy discussed in terms of inches, but I think that refers to the military gps systems.

 

No, it used to be a lot worse that 3 meters! Used to be more like 30 meters, until they turned off "selective availability".

 

Hmmm, I guess that IS the more accurate new version...

"The signals are so accurate that time can be figured to much less than a millionth of a second, velocity can be figured to within a fraction of a mile per hour, and location can be figured to within a few meters. Typical horizontal positioning accuracy for military users is 5 to 10 meters, while for single frequency users is 10 to 20 meters."

(Taken from http://gps.losangeles.af.mil/jpo/gpsoverview.htm)

 

You can get that kind of accuracy with a differential GPS system. The regular satelite constellation is augmented by a ground-based transmitter. Since the ground transmitter location is known exactly, you can get position information relative to it on the order of centimeters. That's how cities use GPS to create digital maps, for instance.

 

We did add a differential system a couple of years ago, for testing at Willow Springs. For racing, it really did not appear to have an edge over our standard GPS system. I tested a version of one of our GPS systems, at Ky Speedway around two years ago, that did provide accuracy down to the centimeters range. We performed that test trying to determine the validity of GPS data for racing/engineering simulation purposes. It was really great data, in the x, y and z direction, but not useful to racing. The system required an hour to accquire the satellites and if satellites were lost (example..parking in a garage) it required another hour to locate the satellites.

We have designed, tested and eventually discarded a number of different systems
in order to arrive at what we now have. We have tried to keep it simple, yet accurate...as we don't want to put a missle in the enemies window, just provide accurate laptime, mapping and speed data for racing.

Later tonight, I will post a screen shot from one of our tests with our professional teams. They have created a page within our software that displays GPS laptimes vs trackside beacon laptimes. It provides a good reference between the two.

Thanks

 

traqmate
anyone tried out a traqmate system?
Heard some folks love them best..

 

We've had two very detailed reports of firsthand experiences with other systems. I'd say this system needs more than "heard some folks love them best" if it wants to join the party.

 

As an autocrosser, I am also interested in data acquisition. I've been using a simple biaxial accelerometer based system for some time: Geez. It produces crude track maps, but it does a good job to show where you didn't fully utilize the car and could have gone faster. However, I've reached the point where it is usually easy to fully utilize the car, and I need to work more on racing lines and car placement.

These GPS systems intrigue me, but I think my positional accuracy requirements are too high. At the level I race at, driving lines are what I most need to analyze. I would like to be able to analyze lines with at least 6"-12" of positional accuracy. Do any of these GPS or GPS/accelerometer systems provide that sort of relative accuracy over, say, a 60 to 90 minute session? What is the best relative accuracy provided during a 60 minute time period?

Steve

 

Hey, I was looking for the feedback from the folks who had used it.

 

Download the trackmate demo software and try it. It is severely lacking. Not even worth considering at this point. In addition, currently they only have plans for inputs with no software functionality to support it.