IMU inverse kinematics

[Zhu Bin]

Dear all:
I want to know the pricinple of the IMU inverse kinematics,when I search the document of this:https://simtk-confluence.stanford.edu:8 ... cs+Works,I find the sentences as follows:
An orientation error is the orientation difference between the experimental IMU and the corresponding model's IMU, quantified by the the angle when representing the orientation difference by an axis-angle representation.

I can't understand what the corresponding model's IMU,I know the experimental IMU is the quartenion data from the IMU sensor,which is included in the orientation.sto file,but how can I get the corresponding model's IMU,anyone can help me?Thanks a lot!

[Ayman Habib]

Hello,

The use of IMU Inverse Kinematics tool involves two steps:
1. IMU Placement: This step "places" IMUs on some of the model segments in a specific orientations. This is done once.
2. IMU Inverse Kinematics: This step solves for the model kinematics that make the IMUs placed in step 1 follow the recorded/desired orientations.

The IMUs placed on the model are what's referred to in the documentation as the model's IMUs.

Hope this helps,
-Ayman

[Zhu Bin]

Thanks Ayman's reply,
it seems that the method of the IMU-IK tool is orientation-based inverse kinematics,I found the related paper:
doi.org/10.1080/10255842.2018.1522532
it seems that the experimental IMU's quaternion data is the "ground truth",and the calibration osim.model's orientation is the initial pose,and we search the relate quaternion to turn the initial pose into the "ground truth",and we use the gradient descent method to search the q,and the q is the IMU-IK's result.is it right?
Thanks a lot
-Zhu Bin