Torso Scaling

[Josée Mallah]

Hello,
I am using the human walking dataset of C. Fukuchi et al. I managed to scale the gait2392 model in a way that the markers match, but I noticed that I am using a scale factor of 1.0 for the torso. In fact, all the markers in the dataset have been placed on the lower body, so I do not have any upper body markers to define the torso.
Do I actually need to scale the torso, especially for further simulation e.g. IK, CMC and metabolic cost? If so, any idea how to accurately estimate a scaling factor for the torso?
Thank you

[Mohammadreza Rezaie]

Hi, CMC requires a full-body model, otherwise you will end up with huge errors in kinematics and dynamics (CMC: Evaluating your Results). I cannot think of an easy way for adjusting the torso mass properties (mass, mass center and inertia) without knowing its accurate scale factors. But you can safely use Static Optimization tool for your (pelvis+lower limb) model to estimate muscles force and activation. In this way, you still need some extra processing as discussed here: viewtopicPhpbb.php?f=91&t=16692&p=0&start=0&view=

[Josée Mallah]

Thank you for your reply. Based on your response and the linked discussions, I decided to keep using the gait2392 model, but remove its torso and connected muscles. I ran scaling using the remaining mass. For some reason I am unable to save scaled models, so I am attaching the original model and the scale file.
What do you think of the result?
I thought it was very good according to RMS and max marker error, but now I am thinking that joint angles might not be quite right, as my IK results seem to have an 'offset' (especially visible at the ankle, which neutral position is -18).
I then ran ID, and results got even more messy - the ankle moment for example does not look right (has the same trend as Fy). I am also attaching some necessary settings file (for force data FP2 is right and FP1 is left). I've also posted further info on this here viewtopicPhpbb.php?f=91&t=17829&p=0&sta ... bd32578942
Thoughts? Any insights are much appreciated.