How does an AHRS compensate for inertial sensors that are not aligned with the airframe?
In this case, the integrated / filtered orientation (i'll call it the sensor orientation) is in the form of a unit quaternion. The difference between airframe orientation and sensor orientation (i'll call that the installation offset) is known and is also in the form of a unit quaternion.
Initially I thought it would be a matter of simply multiplying the inverse of the installation offset by the sensor orientation. Well that doesn't work. The gyroscopes seem to have mixed up axes. Pitching up for example results in a change of heading.
Can anyone help with this?
Cheers, Sam
Replies
That is the problem Samuel; the lack of orthogonality leads to cross coupling and the cross coupling source cannot be easily isolated, ie, when as you see, you pitch you also roll, and may even yaw as well, etc..
I have responded in another post of yours.
Joe
So if you have sensor orientation as p and "offset" as q, wouldn't you want to do p' = qpq-1 ?