The data sheet for the ADIS16480 states a static roll/pitch accuracy of 0.1 degrees. How long does it take the EKF to converge to this accuracy? Are there any requirement regarding movement (angular or linear) of the ADIS16480 in this period?
The short answer is yes, but the accuracy is tied to specific test conditions. When you run this on your platform, we look forward to discussion your observations.The magnetometer influences all three axes but is more influential on the yaw axis, which cannot use accelerometer-based orientation measurements, with respect to gravity.
I hope that helps.
The EKF settles in ~825ms, after the ~500ms start-up time. Here are a couple of things to think about.
So keeping the unit stable for about 2 seconds should be sufficient to have a roll/pitch accuracy of 0.1 degrees then?
Does the magnetometer errors influence the pitch and roll estimates or is it only the heading estimate that depends on the magnetometer measurments?
Retrieving data ...