simtk.org

Hi everyone,

I'm running a simulation with a modified dynamic upper extremity model to derive muscle and joint reaction forces (JRF) acting on the clavicle.
The model contains several coupler constraints and I'm not sure how they influence JRF. Enabling or disabling the constraints during the computation of the JRF results in different JRF's and I don't really know which one is more correct. (See figure attached)
To check where the difference originates from, I checked the bodykinematics acceleration, since the acceleration is necessary to calculate JRF (Equation A, Steele 2011, supplementary).
The accelerations are different when enabling or disabling the constraints, but are in both case unrealistic. From another topic on the forum I read that changing any forces in the simulation will affect the equations of motion and thus change the resulting accelerations calculated.

Does anyone know which accelerations are used to solve for the JRF and how coupler constraints influence calculation of JRF?
https://drive.google.com/open?id=18qC6O ... neubo1xeoP

Thank you very much for your help!
Sanne

In the case of constraints, things are not so straightforward. If I were you I would try to use a simpler model that uses joints for the clavicle complex. Essentially constraints (Lagrange multipliers) bear some of the reaction loads too, and you need to take into account those in the reaction force calculations. However, constraint forces may not be realistic because the constraint Jacobian (lambda = Phi^{-1} tau_con) may be rank deficient. Furthermore, their interpretation as reaction loads depends on the definition of the constraint.