simtk.org

I am trying to replicate Tutorial 1 using my own data. I have a 4x opensim model, a .zml file and a scenario, following the setup for Tutorial 1. I have changed all DOF names to match those in my own model. However, I am getting a bizarre error when loading in the model:

SCONE version 2.0.0.2243 RC 3
SCONE Tutorials and Examples are up-to-date
Successfully initialized OpenSim3 version 3.3-2021-01-28
Successfully initialized OpenSim4 version 4.3-2021-08-04-a68c7148e
Could not find initial state for /jointset/ground_pelvis/pelvis_ty/value with external force of 131.517
Created model Human_SCONE_noMasses; dofs=54 muscles=78 mass=67.0548

If I ignore this issue and evaluate my scenario, then it will freeze on 1% and no longer be responsive, but I can visually see my model disarticulate with the toes flying off into space in SCONE. I assume the issue is related to the above error. I have combed through my opensim model and the .zml file, but all DOF names are the same. I have exported the coordinates from SCONE itself and used that as the initial state - but I still get the same error! I'm at a loss as to where else to check. Is there some other file elsewhere that might be reading in data? From my understanding, I only need to modify the following files to replicate Tutorial 1:
1. the opensim model
2. the .zml file
3. the scenario to read the correct model and .zml

Previously, my DOF was called Pelvis_Ty and the error stated that it couldn't find the DOF 'pelvis_ty' which didn't exist in my model. I changed the DOF name in my model to match the missing DOF (i.e., to be consistent with case sensitivity), and now I get the above error stating that it cannot find the initial state, even though it is listed in my .zml file as 'pelvis_ty = 0'.

Any thoughts?

Thanks so much,
Ashleigh

The error indicates SCONE cannot determine the initial height (which is pelvis_ty by default) of the model with respect to the ground plane. A common cause is that the contact geometries and contact forces are not setup properly in your model. Have you tried running a forward dynamics simulation with your model in OpenSim, to see if it behaves correctly?

You also might want to use "Examples/Standing High Jump - OpenSim4.scone" as a starting point. This is the same as Tutorial 1, but it uses an OpenSim4 model, which should make it easier to compare with your model. If the Coordinates in your model have the same names, then you do not need to modify the .zml file that's used for the initial state.

I hope this helps!

aaaah! I thought it was a problem with the DOF name.

My model was setup with the centre of the pelvis at location 0,0,0 so I set the default value of pelvis_ty to be 'standing' on the ground. But I assume this might be the issue then! I'll investigate that issue. Thanks for the advice!

Thanks for the help, Thomas!

As an update, I ran forward dynamics in opensim with all joints apart from pelvis_ty locked. I placed the model above the ground plane and let gravity pull it down. The model successfully bounced on the ground plane, indicating that my contact spheres and ground plane has been successfully established.

I've played around with height, starting positions, changed the position/radii of the contact spheres etc, but nothing I do seems to work and I still get this error:

Could not find initial state for /jointset/ground_pelvis/pelvis_ty/value with external force of 131.517

Any thoughts?

Thanks so much in advance, and apologies for what it probably just my silly mistake somewhere!!