simtk.org

Hello!

I was woundering, if I can use OpenSIM to estimate residuals. I have not worked much with biomechanics, so my questions can be a little stupid.

What i have as input data is motion (marker coordinates), body parts weights (also moments) and reaction forces. I want to estimate, if the motion is physically plausible, e.g. if the residuals computed by inverse dynamics are very big or tolerable. Can i use OpenSIM GUI to get residuals? Is there a tutorial on how i can do that?

Best regards,
Alina

Hi Alina,

If you do have a model, scaled to match your setup including all the segments and muscles that you want to model, and also some measurements of forces applied on the model from the environment (e.g. ground reaction forces) then you can
- run IK to compute Joint angles.
- Run Inverse Dynamics (ID) to compute the moments/forces required to produce the IK solution. (Both of these steps are described in Tutorial 3 for a gait model).
- At this point you can evaluate the results of ID to see if they are reasonable (for example can be produced by the muscles in your model) or if not then the difference will have to be accounted for as residuals. You can take this further to compute muscle forces as well.

I'm not sure what joint reactions you have, what do they measure or how they were computed, it would be helpful if you can provide more information.

Best of luck,
-Ayman

The joint reaction come from the force plates. And marker position data comes from marker tracking system, but the data is smoothed. I was also woundering, if i can approximate reaction forces myself based on the weight on the person. Is there any possibility to do this or will it be too imprecise?

And how can i find out, if the results can be produced by the muscles. Do you mean to use standard biometrical values to see if the results are reasonable? I thought it can be very imprecise method.

Another thing, when i was going through tutorials for OpenSimi, from the description of ID algortihm I had an impression that one can calculate residuals separatly. Is it not so?