SparkFun Forums

[hoznin2]

Hi,

I would like to add a GPS to my IMU to create an INS. The problem is that I need the GPS measurements relatively accurate for short period of time.
I have searched many webs but I have not found answers to my specific questions. They are:
When I get measured data from the GPS, there is some noise/error. What is a character of that error for a short term (say minutes), in outdoors conditions, several Hertz sampling rate. Is it a pure white noise around a mean (position) or does it have another character (say slow moving of the mean, stairs, sawteeth ...).
How accurate are the measurements? But I am not interested in exact global poistion. Lets say that I have measured some position in the beginning of measurement. How would the measured position change in several minutes.
And what about verical accuracy? It should be worse than horizontal.
What chip is the best for me?

This above are my basic questions. If anyone knows write please or send a link to some pages where I could find answers to this sort of questions.

thanks for help

[olivier_p]

Hi,

The receiver will probably only refresh the position 4-5 times per seconds. To get the most accurate differential position, the distance actually travelled between 2 sampling should be as short as possible to cancel out as much error as possible. Therefore it is more accurate on slow moving objects.

[Krogoth]

Actually, you will get *less * error if you move further between each point. GPS receivers typically have a 5M RMS position error, therefore your position could be anywhere within 5m of where you say it is. Therefore, if you move 1m, your maximum difference in position error is (1 + 5) - (0 - 5) = 11m, or 1000%. If you move 100m, your maximum error is (100 + 5) - (0 - 5) = 110m, or 10%. Remember, if you're talking about accuracy and precision in measurements, they're not the same thing

I suggest you read the application notes on the U-blox site, they contain a wealth of information on GPS and what you can expect from them. Alternatively, buy yourself a GPS and characterise the error yourself.

[NleahciM]

The noise I've seen has not been white noise at all. Often times there will be a fairly constant offest along with a drift. I've seen GPS positions drift by 10s of meters - in that it'll start out in the right position, then with each refresh it'll move a meter or two in the same direction.

However this is without a good signal. If you have a good signal I'm not entirely sure what sort of noise to expect.