Differences between Static Optimization and CMC?

[KATHLEEN BALFOUR]

Hello!

What are the main differences between Static Optimization and CMC? In what scenarios would using one be better than the other?
In which tool can we import EMG?
For example, the project I am working on is using OpenSim to find the knee joint contact loads during ergometer rowing, using EMG data.

Thanks in advance,
Kathleen Balfour

[Dimitar Stanev]

Hi,

What are the main differences between Static Optimization and CMC?

Static optimization (SO) is faster. SO partitions the simulation time at some intervals and evaluates the optimization criterion without considering what is happening before or after (each time step the computations are independent). On the other hand CMC is a dynamic optimization, meaning it takes into account the evolution of the system. CMC solves the problem as follows: 1) produce a set of muscle inputs, 2) integrate and produce trajectory, 3) compare the trajectory with the desired one and 4) adjust the muscle inputs based on the tracking controller.

In what scenarios would using one be better than the other?

You should consider simulation time. CMC can be more accurate, in sense that it can better reproduce the activation patterns. During SO you can choose if you want to approximate the muscle behavior (fast target) or use a better approximation. Also SO, does not consider activation dynamics (delay), because it solves the problem independently (no memory). There is not answer to your question it depends on your problem, but I would go first with SO.

See (Anderson, F. C., Pandy, M. G., "Static and dynamic optimization solutions for gait are practically equivalent", Journal of Biomechanics, 2001)

In which tool can we import EMG?

What you can do if you have your EMG data is to further constraint the muscle activations that are computed by the optimization, because either way you need to solve a redundant problem. I think that you can't constraint SO, but you can apply activation constraints during CMC. So in your case, you probably need to use CMC.

http://simtk-confluence.stanford.edu:80 ... d+with+CMC
http://simtk-confluence.stanford.edu:80 ... d+with+CMC

Best

[KATHLEEN BALFOUR]

Thank you!

[Charles Khazoom]

Hi,

I have also been wondering the difference between SO and CMC, as the faster SO seems to yield similar results to CMC in many cases even if it is neglecting muscle activation dynamics. Why would someone want to consider muscle activation dynamics (thus use CMC) for walking, where the activation dynamics is probably so much faster than the slower, more impotant rigid-body dynamics. In particular, why was CMC prefered to SO in this paper :


Dembia CL, Silder A, Uchida TK, Hicks JL, Delp SL (2017) Simulating ideal assistive devices to reduce the metabolic cost of walking with heavy loads. PLoS ONE 12(7): e0180320. https://doi.org/10.1371/journal.pone.0180320

Thank you for your help !