Static Optimization using Dynamic Upper Extremity Model

[Jonathan Egan]

Hello!

I am attempting to use the Upper Extremity Dynamic Model to determine the muscle force profiles that will produce specific motions of the wrist and I am running into a problem I am hoping someone out there can help me with. The motion I am currently working with is a simple flexion-extension of the wrist (the data I am using to compose the .mot file is theoretical, not collected via motion capture). When I run SO, I receive a message like "time = t Performance= 0.004999999 Constraint violation= 1.#QNAN" at all time points and the output Excel sheet is very odd: the last 14 columns are void of data, despite being labeled with muscle names. Also, the first 10 columns (10 out of the 13 passive force elements, e.g. ligaments, damping force) are filled with "1.#QNAN".

Since SO was not working for me, I attempted to use the ForceReporter under the Analysis tool. I am able to get data from this method, but the forces output are up to 5kN, which seems much too high to be accurate. When I simply use the Plot tool to plot flexion angle vs. total fiber-force for the muscles I'm interested in, the magnitudes of the forces are about 10x less than that output by the ForceReporter.

If anyone could please give me some advice, particularly how I may be able to get SO working, I would greatly appreciate your help. Thank you!

[jimmy d]

It is a problem with either your model, you motion, or both. Hard to say without looking at the all the setup.

[Samuel Hybois]

Hello !
Did you append an actuators .xml file to your musculoskeletal model ?
I noticed that it helps to make static optimization converge (although it does not guarantee the validity of the results, but then you can enhance your model and IK data...)