I’m looking at a log of Pixhawk 6X which has 6 EKFs running. It seems to have a problem with “high accel bias” and switches between the EKF instances.
I’m looking at the different estimators and their accel biases and it looks like only instance 0 and 1 actually have a non-zero bias. Is that expected? And if so, why? Are there different configurations/settings for the different EKF instances? And if so, where can I look up these settings?
Nevermind, as @bresch pointed out, I didn’t actually plot the right data 
.
The plots should look like this:
So ignore my first questions, however, this one remains:
Are there different configurations/settings for the different EKF instances? And if so, where can I look up these settings?
And this is the log by the way: https://review.px4.io/plot_app?log=502f613c-84b3-4405-b8fc-07ac372a3a1b
Currently, all the instances are using the exact same parameters; they are just using different sensor combinations.
