simtk.org

By following the webinar on joint reaction loading, I managed to proceed trough all the steps and conduct some results.
Although comparing these results with literature, it is certain I made some mistakes in the process.

Goal of the study:

I have recorded a motion capture file of a subject with his ground reaction forces before and after a gait retraining program. The objective is now to see which load is applied to the tibial bone during gait.
The motion capture file has been imported in Visual 3D and from there an export to OpenSim is done resulting in the following files:

Visual3d_Scale_ScaleSet.xml - to scale a standard model in OpenSim to my subject
Visual3d_SIMM_grf.mot - ground reaction forces
Visual3d_SIMM_input.mot - kinematic motion files - which as I read on the OpenSim website already contain the inverse kinematics?

Next I opened up a model in OpenSim and scaling it via the setup scale file I've created:

GVB_Scale_ScaleSet.xml

When checking the new features of my model, this seems correct.
Next I've associated the motion file to my model.

Visual3d_SIMM_input.mot
Visual3d_SIMM_grf.mot

Then I've run the inverse dynamic tool with the external forces via the following set up:

ID_SET_UP.xml
external forces: External_loads_setup.xml

Then running the static optimization via:

Static_optim_setup.xml

To then eventually run the Joint reaction analysis via

External_loads_setup.xml

Only taking into account the ankle joint reactions in the parent reference system as I want to calculate the reaction forces on the tibia.My results are however completely off compared to results I found in literature. I think I've followed the webinar in a correct way to solve this problem so I'm a bit lost at this moment. Would it be possible to get some feedback on my solution?
You can find my setup, motion and result files in my shared drive.

https://drive.google.com/drive/folders/ ... sp=sharing

Many thanks in advance and with kind regards,

There are a lot of steps between the input data and getting out joint reaction loads which can make it difficult to find where an issue could be. We usually suggest breaking down the problem into the individual steps along the pipeline to isolate the problem. For instance:
1. How are the errors after the inverse kinematics step? If they are large, you may want to revisit the scaling and IK step to improve those.
2. How do the inverse dynamics moments look compared to the reference data you're comparing to? Issues here again likely point to issues with scaling, IK, or even the data itself.
3. How do the muscle forces/activations from static optimization compare to reference EMG for the motion of interest?

Since joint reaction analysis is just an analysis after all the previous steps, it's typical for issues to be in the previous steps that lead to problems seen in joint reactions analysis.

Hi,

The model you used was gait_2354 which is not strong enough due to the low number of muscles (54). I would suggest you use the Rajagopal model for your research which is actuated with Millar2012EquilibriumMuscle. you can find it here: https://simtk.org/projects/full_body/

But as you use Visual3d for preparing the required motion files for Opensim, based on your visual3d version, you should consider the Rajagopal model before exporting OpenSim motion files. here is a tutorial link for that: http://c-motion.com/v3dwiki/index.php?t ... al_OpenSim
Although it is a little complicated, but it worth. Otherwise, you can manipulate the input file to adjust it for the Rajagopal model.

The default Visual3d_SIMM_input.mot file works well with gait_2392 and Hamner models. They are good too but I suggest you replace the Thelen2003muscles to Millard. You can find it here: https://github.com/opensim-org/opensim- ... 92_Simbody

Moreover, you might like to run RRA before SO or CMC to reduce the residual forces and adjust the model's center of mass. If you did it, don't forget to change the COM in your RRA actoators.xml to the new COM of the scaled model and also set RRA kinematic_q as the kinematic file of external loads.

These can enhance the prediction of the joint reaction analysis tool.

Regards