simtk.org

Hi all,

I am attempting to compare OpenSim 3.3 JRA estimates to hand calculations for a simple model (one joint, 4 muscles). I am getting values that are slightly off (1% or so). My main question is with regards to the accelerations of the bodies used during the Joint Reactions Analysis. How can I find out what these values are? Better yet, is there a tool anyone is aware of that can can be used to help map out exactly what JRA is doing (e.g. lists forces, accelerations, etc. that help determine the reaction forces)

One issue that I have seen many users talk about is the fact that accelerations from BodyKinematics are calculated based on F=ma. Are the acceleration values used during JRA the same ones I should get as an output from BodyKinematics? If so, I am having an additional issue where the results from BodyKinematics change based on whether or not muscle controls are used as an input.

Thanks,
Ryan Byrne

Hi,

You can find more details on the internal implementation here:

https://github.com/simbody/simbody/blob ... em.h#L2257

and some tests:

https://github.com/simbody/simbody/blob ... Forces.cpp

For more accurate testing as mentioned in your post, you will have to implement the process on your own.

What is the comparison that you made and found differences?

Thanks Dimitar,

I will take a look at these resources. The model is a lumbar spine joint undergoing flexion with a moving joint center in the superior-inferior direction. I found differences in super-inferior joint reaction forces, although differences were only small <0.6% when using accelerations calculated by BodyKinematics (without muscle controls included). When muscle controls were included, accelerations from BodyKinematics were much lower and when hand-calculating the superior-inferior joint reaction forces were higher (about 1.5%). The values are pretty close, but the fact that they are different makes me wonder if I'm overlooking a component.

Ryan