Inverse Dynamics Point Force nans

[Kate Mockler]

Hi all,

Relatively new OpenSim user here. I am trying to work out why my resulting inverse dynamics .sto file has 'nan's in the pelvic force and moments columns. I only have force data for the right foot ie data from one force plate. Is this sufficient to compute inverse dynamics? Or do you need forces for both feet? I inputted the ground reaction force as a point force to the right calcn.

Thank you in advance.

[Kate Mockler]

Just incase anyone has a similar problem, here was my solution:

The problem was with my ground reaction force file (the .mot file from the walking trial, exported via Matlab C3DExport.m). At the end of this file, there were 'nan' values instead of numbers. This caused the inverse dynamics output file to have columns full of 'nan's. To fix this, it was necessary to open each ground reaction force file in a text editor (notepad++, atom, sublime text) and do a find and replace. I replaced all the 'nan's with zero. This fixed the problem with the inverse dynamics output file and eliminated the nans.

Hope this helps!

[jimmy d]

If possible, you can delete the entire row of data that contains NaNs. This may be useful when the NaNs exist in a region that is actually necessary for ID (like during stance). The internal spline doesn't function doesn't require linear time points.