simtk.org

Hello,

I am currently working with the Gait2392 model and adapting it to my needs. I had some problems understanding the definition of the joints, especially the hip and knee joint. The hip joint is defined between pelvis_offset and femur_offset, while the knee joint is defined by femur_offset and tibia_offset. Since they represent the proximal and distal ends of the body, they both have a femur_offset frame. So far, so good. For every other joint (except these two), there is a translation offset for one of the two offsets because they can't be in the same place in the body.
Obviously it works and the kinematics seem fine too, but is there a reason why both hip and knee joints have the same femur_offset in (0 0 0)?
As I understand it, the femur_offset for the knee joint must be at (0 -0.44 0), not at (0 0 0). Have I overlooked something?

Thank you for your help.
Kind regards
Johannes

Johannes,

The knee is a little different in this model, because the axis of rotation translates relative to the femur. This translation is coupled to the rotation (i.e. a function of knee_angle) and you can find this in translation1 (x) and translation2 (y). Translation y varies from -0.39 to -0.42 and depends on the knee angle.

They could have indeed defined an offset of (0,-0.44,0) and have the translation vary from 0.05 to 0.02 as a function o of knee angle. This would have been perfectly equivalent (only the total matters), but potentially confusing because the -0.44 number is arbitrary and has no physical meaning.

If you're wondering where this comes from, the gait2392 model is described in Delp et al (1990), which cited Yamaguchi and Zajac (1989) as a source for the coupled rotation-translation in the knee.

Ton van den Bogert

Thank you for your help!