static optimization vs CMC

[Kang Li]

Opensim provides two tools to estimate muscle forces. We have successfully applied them to model running, stairclimbing, walking. However, the results from static optimization and CMC were not really comparable. For example, before heelstrike, the quadiceps forces from Static Optimization usually were very small, CMC provided more reasonable forces. After heelstrike, the quadiceps forces from static optimization were much smoother than the forces from CMC. The force magnitude were also different.
I am wondering which forces should be used for further analysis. Should we question the validity of the static optimization tool of Opensim?

[Ayman Habib]

Hi Kang,

The two methods would not produce the same results because of the different assumptions built into them. CMC performs integration to advance the system through time leading to smooth activation and state profiles at the expense of slow/complicated process. The look ahead window used allows CMC to start activating muscles in anticipation of a future event if it falls within the look-ahead window. Static Optimization on the other hand is a frame-by-frame solver that practically does Inverse Dynamics then solve for muscle activations/forces to produce the Inverse Dynamics solution. It's much faster but has the caveat that the solutions are not smooth, neither are states and the sampling is much sparser than CMC (only computed for specified frames rather than from integration) . You can use it to figure out general trends and/or bounds on activation/forces in a quasi-static sense. Based on your research question you'll have to decide for yourself which one is a better fit for you.

Hope this helps,
-Ayman

[Max Rusch]

Hi Habib,

I posted the files you asked me for static optimization, in the session "features & bugs". Im looking forward.

Thank you very much, Max Rusch

[Ayman Habib]

Hi Max,

I left you a comment asking for the setup file you're using, did you post it? What was the bug number?

Thanks,
-Ayman

[Max Rusch]

Hi Ayman,

I post the files you requested. The Bug number is [#1395] with name: Static optimization tool - Constraint violation problem.

Im looking forward.

Thanks, Max