RRA Mass Adjustments...Inertia?

[Nicos Haralabidis]

Hi all,

I seem to be posting quite often recently, so apologies for spamming the forum!

After receiving the recommendations from RRA (adjust torso COM and segment masses) I have updated my .osim Model file. I then started to think about the Inertial properties, as surely they should change to coincide with the changes made to the mass and/or COM location of a segment? Perhaps I am over thinking it... Any help/suggestions is greatly appreciated.

Kind regards,

Nicos Haralabidis

[Thomas Uchida]

I then started to think about the Inertial properties, as surely they should change to coincide with the changes made to the mass and/or COM location of a segment?

The inertia tensor is updated for all bodies (excluding Ground). Here are the comments in Body::scaleInertialProperties():

Code: Select all

...
// Scale the mass.
...
// Scale the inertia tensor.
// If the mass is zero, then make the inertia tensor zero as well.
// If the X, Y, Z scale factors are equal, then you can scale the
// inertia tensor exactly by the square of the scale factor (and
// possibly by massScaleFactor), since each element in the tensor
// is proportional to the square of one or more dimensional
// measurements. For determining if the scale factors are equal,
// ignore reflections-- look only at the absolute value of the factors.
...
// If the scale factors are not equal, then assume that the segment
// is a cylinder and the inertia is calculated about one end of it.
...

[Nicos Haralabidis]

Hello Tom,

Thanks for your reply and sharing the snippet of code! Am I correct in thinking that this only applies during the scaling stage of a model? and not when making the recommendations following RRA?

I have checked the Rajagopal data repository and it appears that whilst the mass of the segments and Torso COM location changes following RRA, the inertia tensor remains strangely identical between models - to 15/16 decimal places.

Kind regards,

Nicos Haralabidis

[Josh Baxter]

My understanding is that RRA isn't trying to make the segment inertia correct but rather make them dynamically consistent, which can be achieved by changing the mass of each segment.