Applying residuals at pelvis

[Giorgos Giarmatzis]

The movement I am looking at includes a subject sitting on a chair and performing knee extension against an external force applied though a cable at the ankle.
I first run inverse dynamics (ID) with only the external force. Then I apply the residual forces/moments taken from ID solution to the origin of the pelvis as external forces/moments. When I run ID again and although all residuals are expected to be zero, only 4 out of 6 components are. Namely, the three residual force components and the residual moment around Z axis - pelvis tilt. Pelvis list and rotation are not zero.
Why is this happening?

cheers

[jimmy d]

Why are you applying loads at the Pelvis? It doesn't sound like it is necessary for your ID solution, which I assume is for the knee.

[Jack Zhao]

Why are you applying loads at the Pelvis? It doesn't sound like it is necessary for your ID solution, which I assume is for the knee.

Dear James and Giorgos:
This is just what i am also studying and confused. I assume that Giorgos means:
The reaction force from chair to the pelvis is another external force and it is fairly important for simulation, while it cannot be measured experimentally like ground reaction force in gait. (in gait simulation all the external force is only the ground reaction force, no others; hence regularly we measure it using force plate and then ID, SO...)
Here in order to obtain dynamic consistence, we have to calculate chair external force. Hence Giorgos adopted
residual actuator to complement the dynamic lack -- just as another external force. He assumes that since ID has calculated the lacked external force in the form of residual force, then if he input the residual actuator as external force together with the external one from the cable on ankle, the model will be dynamically consistent,
i.e. without or very small residual forces.
I am also modeling one shank and foot model with the same problem blocking me, hence ask Forum for help.
Whether the way of Giorgos to use residual force to stand for unknown external load is correct ?

Thanks & Regards,
Jack

[jimmy d]

No. That is an incorrect interpretation of what is happening. The residuals about the pelvis ARE the forces and moments that the model needs to generate to make it dynamically consistent.

[Kevin Tanghe]

If I'm correct you know
[*] The positions, velocities and accelerations of every body
[*] All external forces, except the external force from the chair.

In this case, you can find the unknown external force by the law of conservation of momentum.

[Jack Zhao]

If I'm correct you know
[*] The positions, velocities and accelerations of every body
[*] All external forces, except the external force from the chair.

In this case, you can find the unknown external force by the law of conservation of momentum.

Dear Kevin:
Yes, thanks. But do you know OpenSim provide such function for conservation of momentum
solution ?
In my opinion, it is much possibility that the OpenSim inverse dynamics may contain the principle (conservation of momentum) and provide the result (chair external force).

Thanks & Regards,
Jack