Removing muscles from a model

[Vahagn Stepanyan]

Hi,

I am trying to remove some muscles from the Hamner2010 full body model. First I tried disabling the muscles from the navigator, but static optimization failed to run. I also tried reducing the max isometric force of the muscles to very low values (1-10), but again static optimization failed. Finally, I manually deleted the muscles from the model file, and static optimization was able to complete. However, the muscle forces for all the remaining muscles were reported as zero at all times.

Does anyone have an idea what could be going wrong? Is there a better way to remove muscles from a model?

Thank you,
Vahagn

[Ayman Habib]

Hi,

Both disabling the muscles and/or removing them from the model file should do the trick and behave identically (if not please let us know as it would be a bug).

Please consult the help page on Static Optimization here for explanation of how it works
http://simtk-confluence.stanford.edu:80 ... timization

Do you have any actuators in your model other than the muscles? If, for example, you have residual or reserve forces/actuators that do all the work then the solution will require very small or zero muscle forces.

Hope this helps,
-Ayman

[Vahagn Stepanyan]

Hi Ayman,

I don't have any other actuators in the model. Is there anything else that could be causing zero muscle forces?

By the way, when all the muscles are left in the model, static optimization produces reasonable results. It is only when I remove some of the muscles, that all muscle forces become zero.

Thank you,
Vahagn

[George Mastorakis]

Hi,

Both disabling the muscles and/or removing them from the model file should do the trick and behave identically (if not please let us know as it would be a bug).

Please consult the help page on Static Optimization here for explanation of how it works
http://simtk-confluence.stanford.edu:80 ... timization

Do you have any actuators in your model other than the muscles? If, for example, you have residual or reserve forces/actuators that do all the work then the solution will require very small or zero muscle forces.

Hope this helps,
-Ayman

Hi,

I'm trying to disable muscles but it seems that after Static Optimisation, there isn't any difference.
This is by using 3DGaitModel2392

Thanks

George

[jimmy d]

Hi George,

I tested isDisabled (OpenSim 3.3) on a muscle and saw computation of activation, but no computation or application of force. The isDisabled mechanism most likely has a bug where one computation (activation) doesn't see the flag while another computation (force) does.

This is a contentious issue because it is unclear what isDisabled means in all situations. For example, if you want isDiabled to completely ignore the muscle then you wouldn't be able to visualize it since the muscle geometry path needs to computed.

Either way, isDiabled is having the correct behavior of not computing or applying forces to the model.

Cheers,
-james

[George Mastorakis]

Hi James,

Thanks for the info. I would expect that when I disable a muscle, or a set of muscles, the model will fail to walk and as a result fall down. This is the type of behaviour I'm trying to achieve. Nevertheless, this does not happen when I disable the muscles.

What if I erase the muscles completely - would that work?

Thanks

George

[jimmy d]

That's not going to happen with Static Optimization or CMC. Static Optimization is an inverse method and CMC is a tracking controller, both compute muscle forces to match an experimental movement.

Once you have those muscle forces, then you can run a Forward Dynamics simulation. Taking muscles out of that simulation would result in a model falling down.

[George Mastorakis]

Great! Actually, I found those examples (Induced Accelerations-based controller for balance), which almost do what I'm after. In the case of 3DGaitModel2392 example, there isn't a ground in the model, so the human body drops all the way down to infinity. Using the models from Induced Accelerations, I have the ground, but when the human model hits the ground, the behaviour is not great, at it goes all the way under it. Please see attached image.

Thanks

[jimmy d]

Im not sure what advice you are asking for here. The original question seems answered, can you please post in a thread?

[George Mastorakis]

Indeed!
Cheers!