RRA setups with actuators in both feet separately - Vertical GRF only

[Stefanie Gassel]

Hello OpenSIM-Team,

I would be very happy, for any help concerning RRA setups with point actuators in the feet (instead of the pelvis). Please find some plots, source files and the 3 sets of GRF setup files attached...

I only got vertical ground reaction forces and the COP location for both legs. Fx, Fz are missing. And the focus during the measurements was on the right leg. So right data and kinematics is a lot more trustable than the left leg (lot less errors expected). So I would like to separate the residuals of both legs to receive trustable muscle activations for the right leg. The idea is also to approximate the missing Fx, Fz GRF components by using the RRA result curves.

I first tried to use the RRA tool with residual actuators (for Fx,Fy,Fz, Mx,My,Mz) in both feet, in addition to the pelvis. However the result is that the curves of all three sets of actuators are exactly the same. The forces are minimized by 1/3 leading to a recommended mass change of about 2 kg (as opposed to 6kg without the actuators in the feet).
The forces are relatively high (>25N) while both the moments and pErr angles are very close to zero. Only pelvis_ty is up to 1.75 cm

When I then tried to use the RRA tool twice, for both feet separately (once with the right leg actuators and once with the 6 actuators for the left leg) I get the exactly same curves for both for lLeg_Fx/Fy/Fz and rLeg_Fx/Fy/Fz. And the proposed mass change, moments and COM location change is exactly 0 in both cases. Is it an error in RRA execution? I would have expected much higher residuals for the left than for the right leg.

- Why are the force and moment curves for both RRA setups exactly the same?
- Should I split the Fy ground reaction forces up? Only using the right part/ left part?
- Could it be useful to add animated control constraints depending on stance/swing phase?
- I had set calcn_l/r as body to be adjusted, though during toe lift off, the toes would be important as well. I hadn't locked subtalar and mtp joints in the osim file as recommended on the RRA troubleshooting page. Would that be helpful?
- Could the 6kg of mass loss (8% of body weight change) be due to missing the horizontal forces (though the motion was quite static, just standing on toes)?

Any help is much appreciated!

Thank you very much!

All the best,
Stefanie

[jimmy d]

Hi, Stefanie-

What results are you trying to get? What is the purpose of this analysis?

I first tried to use the RRA tool with residual actuators (for Fx,Fy,Fz, Mx,My,Mz) in both feet, in addition to the pelvis

Why are you adding 6 point actuators to each foot? That is not what RRA is for. RRA requires one actuator for each coordinate in your model. It doesn't make sense to add another 6 actuators on each foot.

Looking at your files, why do you need to run RRA? if you are after joint torques, you can just run an Inverse Dynamics analysis.

One comment is that it looks like your kinematics are well off. You model markers for the trunk are far away from the where they should be, so it looks like you need to fix that. Another thing is that your right foot tracking doesn't look good either. The foot goes beneath the ground, and the ankle is doing a lot of shaking - which doesn't look right. Could you include the original marker (.trc) files that you used for scaling and IK?

-j