Modeling MVIC task using forward dynamics

[Giorgio Davico]

Dear all,

I would like to simulate a maximum voluntary isometric contraction, i.e. subject sitting on a dynamometer trying to extend one leg, but I am not sure what the best approach is.

Considering that I won't have experimental motion capture data, I think the viable options are (please correct me if I'm wrong):

• simplifying the kinematics, reducing it to a 'static' pose or a prescribed minimal motion, and use this (as input MOT file) to run SO or CMC;

• trying to simulate a more realistic scenario, which involves adding contact elements (e.g., Hunt-Crossley forces) or coordinate constraints to model the human/seat interaction AND run FD simulations to determine kinematics, then SO/CMC.

Since we will be collecting both sEMG data (from the knee extensors) and the force generated by the subject (measured by the dynamometer), I would opt for the second option. The workflow would be: Experimental data --> FD (providing EMG data and external force applied to tibia) --> SO/CMC (providing the kinematics generated via FD). However, defining realistic parameters for the contact elements is not trivial.

Does this make sense or am I over-complicating the problem by adding contact elements?

In both cases, how could I model the connection between tibia and dynamometer (right above the ankle)? I thought of a weld joint (that would also allow to predict the relative contact force, to be compared to the force measured by the load cell) or a weld constraint.

Please note that the aim is to determine whether or not the force generated by a personalised model employing an optimal control startegy (SO) is comparable to the experimental value (dynamometer), given the 'same' realistic joint kinematics.

Thanks for your help!

Giorgio

[Carmichael Ong]

You could also avoid modeling any contact by using a CoordinateActuator (https://simtk.org/api_docs/opensim/api_ ... uator.html) and applying the torque directly to the joint as measured by the dynamometer.

[Giorgio Davico]

Thanks Carmichael!

If I got it right, you are suggesting not to introduce a weld joint and a weld constraint to model the cuff, but rather to apply the measured torque directly to the knee joint using a CoordinateActuator. We will give it a go.

What if we want to use FD simulations to validate our personalised models as well as SO/CMC results, by providing experimental sEMG data and predicting the output force generated at the cuff level (in order to compare predicted and measured forces/torques)? Would, to this end, the use of a weld joint and a weld constraint be the most appropriate approach to model the tibia-cuff 'interaction' (i.e. to prevent tibial motion and predict the extension force)?

Thanks again for your help.

Giorgio

[Carmichael Ong]

I think either would work, and it might be worth trying both to see what works best for your workflow. The drawback to using a weld joint or constraint is that it can be harder to use than just applying some joint torques. You should definitely whatever the end goal of your project is to decide whether or not the more "realistic" application of a weld joint or constraint adds any value versus just adding a joint torque.

[Giorgio Davico]

We will try both approaches and see what works best for us.

Thanks again for your help and suggestions.

Giorgio

[解 浩]

Hello

I have a question about how to simulate isometric contraction use the tool CMC.What's the column of desired kinematics's file .

I hope accquire some suggestiion.Thanks for help.