Angular momentum around contact points

[jfslin]

Angular momentum around contact points

[pariterre]

Dear @jfslin

1. The short answer is a "segment" in biorbd is translated into N-Degrees-of-freedom bodies into rbdl. Meaning that you always get as many Bodies as you have degrees of freedom. I therefore suspect you have segments that declares more than one translations or rotations. To get the "segment" value, you simply take the last one for that particular segment. For instance, if your model has 3 segments (all of them declaring 2 rotations), then the AM of the second segment is the 4th row (second being the values for the 1st segment and last being the values for the last segment). This is a bit confusing, but there is no easy workaround for this as sometimes you would need intermediate values.

2. There is no internal way to do that, but you can of course recompute the AM to be evaluated anywhere you want using the parallel axis theorem.

Cheers!

[jfslin]

@pariterre Thank you for the quick response!

1. Ah:

mdl = ik.model
mdl.nbQ()
Out[3]: 47

There, that's the 47. Okay. So "segment" in this case is referring to joints/reference frames etc., while nbSegments() gives me the number of rigid bodies that makes up my model.

2. Thanks for including the bit about the parallel axis theorem. I was wondering if that was sufficient. So to recap, I should run CalcSegmentsAngularMomentum, keep only the vectors that are at the end of the body's "chain" for the cases of concentric joints, then use CoMbySegment and the segment masses to move all the individual AMs to the contact point in question via parallel axis theorem.

Perfect. I will get on that. Thank you.

[pariterre]

Yup, sounds good, the only thing about 2) is that you probably willl have to rotate the AM after transporting, unless you want it aligned with the global reference frame!

[jfslin]

Noted. Thanks a lot for your help.