Error caused by WeldConstraint

[Catarina Ribeiro]

Hello,

I have inserted an AFO on the gait model 2354 and defined the joint. To keep the AFO attached to the model, I defined WeldConstraints "Leg" (between the AFO cuff and the model's leg) and "Foot" (between the AFO foot and the model's foot). When the Foot constraint is enabled, the foot of the model tilts as you can see in the figure appended.

Error Forum.PNG (355.12 KiB) Viewed 392 times

This only happens when the Foot constraint is enabled, otherwise the foot is normal, but the AFO foot is not attached.
Can someone help me with this?

Thanks in advance,
Catarina Ribeiro

[Ton van den Bogert]

I am not sure if this caused your problem, but with the two weld joints, you have overconstrained the system, and this makes it impossible to solve or simulate the kinematics. The foot relative to shank only has two degrees of freedom in the gait2354 model, and the welding removes 6, resulting in minus 4 degrees of freedom. Overconstraining is never a good idea, even if the software is smart enough to detect it and remove some of the constraints.

There are several alternatives to model the effect of the AFO:

• If you only need to eliminate ankle motion, you can lock the two degrees of freedom in each ankle and not any other bodies or joints.

• If you want the geometry of the AFO in the model (for visualization, or because it has ground contact elements on it), you can lock the joints and weld the AFO to the foot only. The disadvantage of this is that there is no force between AFO and shank. If this force is not important to you, this solution will be fine.

• If you want realistic estimates of forces between the AFO and the foot and shank, bushing or spring elements can be used instead of weld joints. But you have to design those bushing elements carefully so they have a realistic load-displacement relationship.

Ton van den Bogert

[Catarina Ribeiro]

Thank you for your fast answer first of all.
I wanted to assess the alignment between the AFO and the model's foot by analysing the ankle moment during walking with/without the AFO with the self-aligning mechanism. For that, I wanted to use the Inverse Kinematics tool.
I don't think I need the force between the AFO and the shank for this, am I wrong to assume this?

Thanks,

Catarina Ribeiro

[Ton van den Bogert]

This depends on how you define "ankle moment". If you need the total ankle moment (from human and AFO combined), you don't need to model the connection between AFO and shank. It won't affect the total.

If you want to know how the AFO affects the ankle moment generated by the human (muscles and ligaments), then you need to model the connection between AFO and shang and foot. Spring or bushing elements would be appropriate in that case.

Ton

[Catarina Ribeiro]

I wanted to know the effect of the AFO on ankle moment. However, I only have access to the gait2354_Scale_MarkerSet.xml and subject01_walk1.trc files in the gait2354_Simbody folder. The ankle moment I am talking about is the one in the .sto files generated by the Inverse Dynamics tool.
Can you enlighten me of whether this ankle moment corresponds to the total ankle moment or to the ankle moment generated by the human?

Thanks

Catarina

[Ton van den Bogert]

The inverse dynamics tool would only give you the total ankle moment (AFO + muscles).

I am not familiar with the AFO research literature, but I believe that there have been studies where they try to separate the total moment into the AFO and muscle contributions. I suggest looking for those studies to see how they did that.