simtk.org

Hi all,

I'm trying to compare old results with some new obtained with opensim and for that I need to obtain the "net joint reaction force", I mean the net forces I would obtain from Inverse Dynamics, without considering the individual muscle forces.

Is it possible?

Thanks
Maria Isabel

Maria,
Yes, that is possible. OpenSim has an in-built Inverse Dynamics tool. Below is a link that describes the Inverse Dynamics tool in OpenSim.
http://simtk-confluence.stanford.edu:80 ... e+Dynamics

cheers,
-james

Hi James,

Thanks!

but the problem is that the Inverse Dynamics tool does not provide the forces I'm looking for (or maybe I don't know how to get them). It gives me the net joint torques at each joint but not the net joint reaction forces.
To get the Reaction Forces I use the Joint Reaction analyses but this analysis account for muscle forces (as far as I understand the forces given by this tool are the net forces that act at the joint and can not be modeled by muscles and constrain forces and, but the ones that can be obtained with the ID are the net forces acting at the joint, including muscle forces, which are not modeled in the ID problem).

I'd like to know if there is a way to acess those forces via Inverse Dynamics in order to compare with my previous results. Do you know how to get them?

jimmy wrote:Maria,
Yes, that is possible. OpenSim has an in-built Inverse Dynamics tool. Below is a link that describes the Inverse Dynamics tool in OpenSim.
http://simtk-confluence.stanford.edu:80 ... e+Dynamics

cheers,
-james

Maria,

Sorry about the confusion. You can get the net joint moments by stripping all the muscles off the model and driving it by torque actuators. The ouputs of that can be used as the inputs for the joint reaction analysis.

Hope that helps,
-james

Hi James,

Some days ago a tried something in this direction but I don't know if the procedures are right.
What I tried was to use de analyze tool with the same inputs as the SO but using the option "Replace models force set" to replace the muscles for torques (and point for the translational DOF) actuators acting at each DOF, and then used the inverse dynamics results as the "force_file" in the Joint Reaction analysis. But the OpenSim (also from command line) crashes .
Is this the right way of doing ?

thank you
Bel

Hi ,

Thanks to James's help I could figure how to obtain the Net Reaction Forces (the one resulting from the Inverse Dynamic, not from static Optimization analysis), and as some details may not be so obvious sometimes, I'll post the procedures I followed in more details:

Basically I followed the same procedure as Matt DeMers in the webinar. The differences are two, as James said:

1) you have to replace the muscle actuators in the model (by choosing the option "Replaces model's force set" in the "Actuator and external Load" tab ) by another set of actuators that will drive the coordinates of the model.

2) Choose the generalized forces from ID solution ( inverse_dynamic.sto, instead of muscles forces from static optimization ) as the input for the "forces_file" option in the Joint Reaction analisys.

For me the tricky was to create the set of actuators:

I created my new set of actuators based on an RRA actuator file from an opensim example. This set consisted of Coordinate Actuators for all rotational DoF of the model (including the ones for the pelvis) and Point Actuators for the pelvis translational DoF. I named the actuators with the same name as the generalized forces were displayed at the inverse_dynamic.sto file. And It stoped crashing.

Hope this post helps saving someone's time in the future.
Bel

Glad it worked!
-james

Hi, this is a rather old post. I did exactly what mbebelveras did. Although, the JRA did not crash, the JR results are different to using SO as force file (smaller and looks similar at the lower limb joints). Perhaps, someone can comment on this.