simtk.org

I'm using a full body msk model (Hamner., 2010) and have used my own motion capture data to run build an inverse kinematic model in OpenSim. The model appears to be replicating the subjects movement relatively well except that there appears to be sudden and unnatural pelvic movements i.e. large and sudden variations in pelvis pitch. My immediate thought is that this may be due to the large marker error but the experimental markers are not visible in the model, so it is difficult to tell in the direction of the marker error. Any other thoughts?

When there are large, sudden changes in body coordinates it is typically because experimental markers have dropped out or the marker labels have flipped. I would suggest checking the experimental data, filling any gaps, and double checking the labels to make sure that there isn't any label flipping.