Problem of rotated model after scaling

[Sylvia Wan]

Hi,

I have modified a XML file which has markers that correspond to my experimental data and that they match the Lower Limb Model (Delp 1990). I also have a trc file for a subject’s neutral position and another trc file for a walking trail. These trc files are facing the same direction as the model when viewed using Preview Motion Data. I tried to offset the model to collide with the blue trc markers and most markers match well except the subject’s feet are more laterally placed than the model.

Using the scaling tool to apply the xml file to the subject data section and the neutral position trc file to both the Scale Model and Adjust Model Markers section caused the model to rotate 90o while the experimental markers stay in the upright position. Also, the ankles appeared to be dorsiflexed. I have read through the forum thread 1245…4228 about scaled model orientation and can ensure that the trc files are in the same coordinate system as OpenSim as seen in the Preview Motion Data.

I wish to resolve this problem of rotated model after scaling. Can anybody help?

Thank you very much.

Regards,
Sylvia

[Andrew Lewis]

Hi Sylvia,
I had a similar problem with a crumpled skeleton after scaling with the BoLegs1793.osim model. The problem turned out to be that the pelvis only had rotational degrees of freedom and not translational. My *.TRC file data was ok but there was a translational offset which the model wasn't able to accommodate and so the model was twisted.

I don't know if this is your problem, but it might be worth checking your degrees of freedom of the pelvis and adding 3 translational degrees of freedom if they are missing (I copied the code from the full body model).

Hope that helps,
Andy

[Sylvia Wan]

Dear Andy,

Thank you so much for your reply. I really appreciated it.

I have checked my model's osim file. 3 axes for translations for the pelvis are defined as in the UpperLowerBody model but the inertia values for xx, yy and zz are zero which is different from the full body model.

I have tried to replace the paragraph associated to the pelvis with the code in the full body model. However, when I tried to loaded the model with the new osim file,OpenSim crashed and shut down.

Was there any problem when you copy the code from the UpperLowerBody model? Did you have to change anything else in your model's osim file to accomodate the change you have made to the pelvis part?

Many thanks Andy!

Sylvia

[Andrew Lewis]

Hi Sylvia,
It was the BothLegs.osim model (one of the example models supplied with OpenSim) that I edited. I copied the pelvis_tx,ty,tz coordinates from another model (gait2392). This involved inserting the <coordinate> and matching <TransformAxis> code into the Pelvis section of the model's BodySet. I didn't have to alter any other bits of code. I didn't spot that the inertia values were zero, but you are correct, they are also zero in BothLegs.osim. I'm not sure if this is a problem. I didn't end up using this model in the end as I decided that I needed a head-arms-trunk segment and so ended up using gait2392_simbody.osim instead.
Sorry not to be anymore help.
Andy