My current project is about dynamic load during walking in children with angular deformities.
We use the Lower Extremity MoCap model provided in the AMMR v2.2.0 and we utilize the TLEM 2.0 version. Additionally we use the AnyKnee Model as AddOn.
When using the 3E-Hill Modell, many muscles are overloaded, but that’s not that “big” problem.
My focus is on the KneeContactForces. We have the problem, that the forces at the knee are really high and inexplicable peaks appear, especially during the stance phase.
Not with all persons/Models this problem occurs however with most. Have you perhaps had these problems before and know how to deal with them? Probably it is also because of the high muscle activity?
If the simple knee model is used, there are no peaks.
Below are two examples of how the knee contact force looks like, not really physiological.
Concerning the AnyKnee model are you using with the fixed or moving axis?
Do you see the same problem if you are using constant strength muscles?
Concerning the hill muscle models, please ensure to run the calibration study before running inverse analysis otherwise strengths will be wrong which may lead to overactivations, which again in principle could lead to some fluctuations.
Please try to look at the accelerations in the model so the rDDot values of e.g. thigh and shank, if you see fluctuations there the issue is kinematically related.
In these models I am not able to reproduce the fluctuations you see in the knee, please describe how your model differs from these, this would help me understand the problem.
I am 99% sure the high tolerance is the reason, in the study you can plot the kinerror, if it fluctuates similarly this is the reason.
I would try to bring back the tolerance to its default value, and solve the kinematic problem.
If the problem occurs in the first frame the solution can be to simply provide a better starting guess on the initial position of the human model.
If the error occurs later in the analysis you need to look what can be the cause in more detail, if the model is marker driven maybe there is a marker drop out.
If the problem can not be solved by providing a better starting guess, please describe how your model is being driven.
thanks for your suggestions.
I run the Analysis with the simple Model (1) and the AnyKneeModel (2).
Here you can see the different results regarding the KneeContact forces.
Our Model is marker driven but I am sure, that there is no marker drop out.
We run the Model with ~30 different C3d files (from 30 different Persons) and with almost all files the same problems arise (Peaks within the KneeContact forces).
Please try to run with default kinematic limit, from your plots it may not look like this is the problem, but i would like to be able to rule this out as the cause of the problem.
If the model does not run with the default kinlimit, this needs to be fixed to be able to move forward.
Please let me know if the model fails in the first frame or if it happens on the way in the analysis?
If it fails in the first frame the initial positioning of the model needs to be improved, if it happens on the way it can relate to the maker configuration.
Pease describe your maker configuration a bit.
Are the model able to finish the parameter optimization?
The kinematic analysis failes in time step 57: "Position analysis is not completed."
So we use the PiG marker set for the lower limb and additionally we have the clavicula, C7 and T10. We only use one marker for the sacrum instead of the right and left anterior superior iliac spine.
Yes the models are able to finish the parameter optimizsation. Nevertheless sometimes it takes time.
