Model rotating / translating during CMC

[Rahul Goel]

Hi All
I also experienced the issues raised here
https://simtk.org/forum/message.php?msg_id=4480

By playing around with starting time/adjust kinematics/lock-unlock mtp or subtalar joints/slow fast optimization, I atleast get CMC to run for some times, but now I experience another issue that my model seems to be rotating about the pelvis or/and gets laterally displaced while CMC is running. I dont think I can trust the results?

Has any one else experienced this? Any pointers?

Rahul
rahulg@mit.edu

[Rahul Goel]

So I observed that after CMC, the 'pelvis_tilt' angle is going out of bound (-90 to 90 deg). Even though it is clamped, it goes upto -213 deg, could it be a result of any singularity?
Any pointers would be helpful.
Thanks
Rahul

[Ayman Habib]

Rahul,

The bounds on coordinates are not enforced during dynamic simulation (to be physically sensible, a force producing element should be attached to each coordinate/bound to model this behavior).

If you believe your model is capable of tracking the motion then I'd suspect something wrong with the external force application or that RRA was not run so there are high residuals that are leading to the non-physical simulation. To rule out the first issue Id run InverseDynamics and see if the results seem reasonable before attempting CMC.

Hope this helps,
-Ayman