Upper Extremity Modeling

[Matthew Edwards]

Hello,

I am new to Opensim, and I am trying to model the upper extremity while doing a curl using the Arm26 model. Specifically, I want to see the different joint torques and muscle forces with respect to different external loads applied to the hand. I have Vicon motion capture data while doing the curl. Basically I am taking motion capture data of the exercise and I want to calculate the muscle forces to produce that specific movement with an external load applied. However, I am confused as to what is the best tool for me to use in order to calculate the moments created by the muscle forces (and the muscle forces themselves) in order to move the model at the correct speed with the applied external load.

To begin, my inverse dynamics solution is way off. I have read that this is due to dynamic inconsistency. Basically, my model can not produce the correct force in order to match the acceleration of the IK solution. Am I correct? What is my best plan of action?

I have tried using static optimization, filtering the IK results, and locking the shoulder in order to simplify the model even further, but I still get bad results.

Any suggestions will be greatly appreciated.

[jimmy d]

Hi Mathew,

ID calculates the joint torques given the motion & model & external loads. Its true there will be residual error due to errors in kinematics and model mass/inertia, but these have little to do with what you are observing.

OpenSim is distributed with an example arm model and simulation pipeline. have you gone through this to see how it matches with your work?

-james

[Matthew Edwards]

Thank you for the reply. Yes and I have now successfully scaled and ran IK and ID with success with my own motion capture data on the arm26 model. However, I would like to use the Upper Extremity Dynamic Model. I have tried running the same trial on the Dynamic model with no success. Even after locking all joints except the elbow_flexion and shoulder_rot, SO will not run with accuracy. I am getting very high constraint violations around 3000 when running SO. The constraint violations are in multiple joints and it is also saying the model is too weak.

Also, the ID outputs an elbow flexion joint torque curve different from the arm26 model. My elbow flexion joint torque curve in the arm 26 model looks more like a parabola with no external loads applied (which makes sense to me) where the Dynamics model just continuously rises to about 8 Nm with no external loads applied. This is too high compared to the 4 Nm I got with the arm26. Also, the elbow moment in the arm26 model makes sense when compared to the muscle forces and moment arms that SO calculates. Any suggestions?

I was thinking about trying to run RRA, but I am having some trouble writing the .xml files for the actuators and tracking tasks being as I am new to Opensim. Do you think this would be a good choice? Or should I try something else?

Thank you for the help.

[Matthew Edwards]

While using the Upper Dynamic Extremity Model, I have gotten the constraint violation value down to around 0 with an external load applied. However, while I am doing a curl, the results say that the tricep is producing a force, around 40 newtons while doing this motion. The bicep muscles are producing 60 newtons. I do not understand why I am getting co-contraction.