simtk.org

Hi. I'm working on running simulation on patients that are walking on a decline treadmill (Bertec duel-belt instrumented treadmill) and having some challenges with extreme muscle forces (i.e. rectus femoris and soleus are >10x BW) after running through both Static Optimization and Computed Muscle Control. I do not have EMG data, but do have literature with similar methods for reference and the magnitude of muscle force is grossly different.

I'm exporting the analog and trajectory data from V3D to OpenSim. Initially, I performed some quality checks by computed inverse dynamics (sagittal hip, knee, ankle moments) between V3D and OpenSim, and the signal and magnitude are nearly spot on. I followed this by running two iterations of RRA to get my residuals down to the "GOOD" quality, based on the wiki, before running either Static Optimization or CMC. I'm using the Kinematics_q files for both the Desired Kinematics for Input and in the External Loads files and the original vGRF file exported from V3D. I also adjusted the Actuators file for the pelvis COM adjusts based on the RRA analysis and commented out the subtalar and MTP joints to be consistent with the .osim model.

I've performed both with and without the subtalar and MTP joints constrained and still getting extreme muscle forces. However, the solutions magnitude of muscle forces are similar between both Static Optimization and CMC. I'm in the learning phases of OpenSim, so any recommendations would be greatly appreciated. Thanks for the time.

JC

My first guess would have been that there was an issue with the data you imported to OpenSim. However, since you checked your joint moments from ID vs the literature that doesn't sound like it is the case.

A couple ideas:

1) Are you using one of the default OpenSim models? Many times if you enable extra degrees of freedom (such as a 3 DOF ankle joint, unlocked MTP joint etc) the muscle paths in the model have small moment arms about these DOFs and therefore you need large muscle forces to generate the measured moments.

2) Add a reserve actuator to each Coordinate, and make them strong (ie max force 1000). Keep rerunning simulations while progressively lowering the strength of the reserves one by one to identify if there is a specific DOF that is causing issues.

I would just stick to Static Optimization for troubleshooting. It solves way faster and if you arn't getting a reasonable solution with SO for walking, you definitely won't get a good solution with CMC.