Whether CMC ignore tendon compliance?

[Jack Zhao]

Dear friends:
Whether CMC also ignores tendon compliance as Static Optimization ?

As we know, tendon slack length is only needed when the tendon force (also is the muscle-tendon actuator force) is being computed. i.e. The fiber length and muscle-tendon length can be obtained from kinematics data, then we can get tendon length to calculate tendon force with tendon slack length.

But Static Optimization ignore tendon compliance, and CMC contain Static Optimization inside.
SO and CMC only calculate muscle-tendon force through active (or in CMC together with passive) fiber force (with only activation unknown) basing on cost function, not via tendon compliance property (using tendon slack length).

While I still saw the instructions in "Getting Started with CMC" (http://simtk-confluence.stanford.edu:80 ... d+with+CMC) that, CMC still need tendon slack length which is even quite sensitive to the result.

Could you please help explain it?


Thanks & Regards,
Jack

[Ross Miller]

Hi Jack,

I couldn't find a definitive answer in the documentation, but the original paper on CMC for gait (Thelen & Anderson, 2006) says that the muscle model included tendon elasticity. As far as I know, CMC uses the full formulation of whatever muscle model you have included in your model.

I say that because if CMC ignored tendon elasticity, then a CMC solution would not be able to control a forward dynamics simulation with elastic tendons, and CMC can do this very well.

Even if the tendon is rigid, the slack length will still be used to compute the fiber length (Lfiber = Lmuscle - Lslack).

Hope this helps,
Ross

[Adrian Lai]

Hi Jack,

In my experience, CMC does include tendon compliance for Millard2012EquilibriumMuscle muscle models. CMC takes into account both the length AND velocity of the tendon component when computing muscle forces (and hence muscle excitations) to drive the model towards the desired kinematics.

Also, like Ross' said, tendon slack length is required for both SO and CMC to compute muscle fibre lengths from MTU lengths but that does not require the tendon to be compliant.

Hope that helps!

Adrian

[Jack Zhao]

Dear Ross:
Thanks a lot! Indeed the tendon slack length is still needed for fiber calculation.

Cheers
Jack

Hi Jack,

I couldn't find a definitive answer in the documentation, but the original paper on CMC for gait (Thelen & Anderson, 2006) says that the muscle model included tendon elasticity. As far as I know, CMC uses the full formulation of whatever muscle model you have included in your model.

I say that because if CMC ignored tendon elasticity, then a CMC solution would not be able to control a forward dynamics simulation with elastic tendons, and CMC can do this very well.

Even if the tendon is rigid, the slack length will still be used to compute the fiber length (Lfiber = Lmuscle - Lslack).

Hope this helps,
Ross

[Jack Zhao]

Dear Adrian:
Thanks a lot! It really helps.

Cheers
Jack

Hi Jack,

In my experience, CMC does include tendon compliance for Millard2012EquilibriumMuscle muscle models. CMC takes into account both the length AND velocity of the tendon component when computing muscle forces (and hence muscle excitations) to drive the model towards the desired kinematics.

Also, like Ross' said, tendon slack length is required for both SO and CMC to compute muscle fibre lengths from MTU lengths but that does not require the tendon to be compliant.

Hope that helps!

Adrian