GPS altitude

cool! i would love to know my altitude.

 

What are the PID parameters for GPS ALTITUDE? That would be interesting to monitor.

 

My Garmin tells my altitude when I touch the vehicle icon, but I've never found it to be very accurate. It's usually off by 50m or more. My Galaxy Note 2 has a barometer to help compensate for such inaccuracies, but it's not been much of an improvement so far. The app I use let's me dial in the barometric pressure, which I get from a weather app. But it also seems to be off. I think there needs to be an overhaul of the GPS system to improve altitude accuracy.

 

is it actually an altimeter, or just based on google maps?

 

It's based on triangulation from satellites, which it notoriously inaccurate.

 

i think the army gave us the bad satellites and kept the good ones, which makes sense i suppose.

 

Very true. The U.S. GPS system is optimized for accuracy in the "X-Y" plane, not the "Z" plane, at least in the public mode we have access to. The government and military has access to a more-accurate augmentation system that can locate in 3-D space to millimeter accuracy: GPS.gov: Augmentation Systems .

 

You point is well taken, but I think you must also consider the ability to time-average the GPS altitude estimates in any geostationary use. Averaging 100s of estimates over time will give far more accurate results.
Estimating alt while driving is harder and while flying harder still as the situation is changing every second, sometimes quite dramatically. I have been evaluating the altitude estimates of a Bad Elf GPS Pro unit as some pilots use it for flight purposes. So far in "automobile" mode the estimates are unstable enough for me to not want to use it in an aircraft for precision approaches without a LOT more careful testing.

 

No BS here. I've used GPS equipment which can determine a position within 4 inches, but it cost $10,000 or more. It uses a number of algorithms for error-correction, and it took about 20 minutes of processing to achieve that accuracy. The same company (Trimble) offered a lower-cost model ($5000) could achieve about 8" accuracy with proprietary PC software which downloaded correction data from a number of base stations, with post-processing.

Instead of calling "BS," I suggest you research the topic a bit. Consumer-grade GPS equipment can't compare with that used in surveying.

 

I think that would be true if the GPS receiver were a point in space surrounded by satellites scattered evenly in the X, Y, and Z directions. But an automobile GPS must cope with a hemisphere of available satellites at best, usually narrower than that, so I believe there is more uncertainty in the Z direction (alt), especially if one operates a car in an urban canyon. It is also quite possible that the solution algorithm can be adjusted to assume things like "only small changes in altitude per second" when solving for the X and Y locations.
The Bad Elf's "aviation" mode setting (software selectable) may more evenly emphasize altitude along with latitude and longitude , but I cannot say for sure until I beg, buy, or borrow an airplane and go flying with it in that mode . But very important to do so before relying on its alt info on a dark and stormy night!

 

Altitude Accuracy

 

As much as I understand GPS receivers, the position is triangulated in orthogonal space (X-Y-Z). Those coordinates are then converted to spheric space (lat-long-elev) within GPS receiver. For the purpose of presentation only.

Given X-Y-Z position of Earth center and X-Y-Z position of GPS receiver, latitude and longitude angles can be calculated quite straightforward by GPS receiver.

However, calculating the elevation is a whole different story. Earth is not a perfect sphere. I remember, I've read somewhere, in order to approximate sea level elevation around the globe one would need to use a 2D polynomial of at least 28th order, which is an overkill for commercial GPS receivers to perform in real time.
Since memory is cheap now-a-days, one would expect to have this problem solved with 2D lookup table stored in ROM by now.

The end result is, the elevation is probably precise about the same as latitude and longitude, but it is not accurate. You just need to determine the offset for your area and your GPS receiver model manually.

For calculating elevation difference within 500mi radius you should be fine using your GPS "altitude" measurement.

 

Since when does anyone get surveying-grade "proper equipment" in a production car?