JointReaction vs InverseDynamics

[Mohammadreza Rezaie]

However, in the graph you showed, it was just ID vs JRA moments, so it was the exact same model, correct?

Same scaled model, IK output, and GRF.
This was the Rajagopal model and the associated sample walking trial. I also tested Gait2354 model and data, there was the same mismatch.

look at the resultant moment from both (i.e., sqrt(Mx^2 + My^2 + Mz^2)

They are quite similar. The RMS difference was only 0.25.

ResultantHip.png (44.59 KiB) Viewed 242 times

Regards,
MrR

[Jacob J. Banks]

Oh, those are the same (maybe some filtering and/or weird little coordinate differences that don't align)! So it is just a mater of which coordinate system you are using.
Are you looking at the child or the parent frame for your JRA? Which ever you are not, should be in the same coordinates as the ID result (my assumption is the parent should be the same).
Jake Banks

[Mohammadreza Rezaie]

Hi,

I tested all possible options of child, parent, and ground for <express_in_frame> in JointReaction class:

Code: Select all

<JointReaction name="default">
 	<on>true</on>
 	<start_time>-Inf</start_time>
 	<end_time>Inf</end_time>
 	<step_interval>1</step_interval>
 	<in_degrees>true</in_degrees>
 	<forces_file />
 	<joint_names> ALL</joint_names>
 	<apply_on_bodies> child</apply_on_bodies>
 	<express_in_frame> ground</express_in_frame>
 </JointReaction>

For <apply_on_bodies>child</apply_on_bodies>:
<express_in_frame>child</express_in_frame>:

CC.png (33.48 KiB) Viewed 211 times

Only hip rotation and MY are the same.

<express_in_frame>parent</express_in_frame>:

CP.png (30.74 KiB) Viewed 211 times

Only hip flexion and MZ are the same.

None was similar for <express_in_frame>ground</express_in_frame>.
The results of <apply_on_bodies>parent</apply_on_bodies> were similar to the above but negated.

If the frame is the source of the mismatch, I think there must be a new option to express the forces and moments in joint coordinate system. e.g:
<express_in_frame>joint</express_in_frame>
Then the output will be identical to ID. Am I right?

Regards,
MrR