simtk.org

Posted: **Thu Jul 16, 2015 7:40 am**

Hi,

I've been working on some code to calculate Normalized Fiber Lengths throughout a walking cycle. I've been able to set up the model to run through the motion by updating the coordinates, but whenever I call either the fiber length or normalized fiber length, a value of zero is always returned. Here's the loop that updates the coordinates:
data is matrix that contains all of the coordinates and their values for each time

Code: Select all

s = myModel.initSystem();
for ii = 1:myModel.getMuscles().getSize
   myForce = myModel.getMuscles().get(ii - 1);
   myMuscle = Muscle.safeDownCast(myForce);
   display(char(myMuscle));
   for k = 1: nFrames
       for j = 1:15
           updCoord = myModel.updCoordinateSet.get(j-1);
           updCoord.setValue(s, data(k,j)/180*2/pi);
           updCoord.setSpeedValue(s,0);
       end
       storageData(k,:) = myMuscle.getFiberLength(s);
   end
end

And when I call for the lengths this is an example of what I get:
>> myMuscle

myMuscle =

addbrev_r
>> myMuscle.getNormalizedFiberLength(s)

ans =

0
>> myMuscle.getFiberLength(s)

ans =

0

If I allow the code to run it will fill up the entire matrix with zeros for all of the muscles. I know the fiber length data is in the model; I can plot it using the GUI. Did I accidentally skip over a setting as I was writing the code?

Thanks,
Colton

EDIT: From just the initial state, I was able to return normalized fiber lengths for muscles in the torso (of which I am not concerned). These muscles are fixed because I am using the pelvis as the origin in a lower body model. Does the fact that I am changing the coordinates mean that the fiber length is lost?

Posted: **Mon Jul 20, 2015 10:03 am**

Try downcasting to the Thelen type muscle model

Code: Select all

% Get the muscleType of that myForce 
muscleType = char(myForce.getConcreteClassName);
% Get a reference to the concrete muscle class in the model
eval(['myMuscle =' muscleType '.safeDownCast(myForce);'])

Always equilibrate the muscles

Code: Select all

% Set the activation and fiber length
myMuscle.setActivation( state, 1 )

% Equilibrate the forces from the activation 
myModel.equilibrateMuscles( state )

storageData(k,:) = myMuscle.getFiberLength(s);

Posted: **Tue Jul 21, 2015 11:55 am**

Thanks for the help again James, but whenever I try to downcast the muscle type, I get this error.

Code: Select all

Undefined variable "Schutte1993Muscle_Deprecated" or class
"Schutte1993Muscle_Deprecated.safeDownCast".

Error in getNFL (line 63)
   eval(['myMuscle =' muscleType '.safeDownCast(myForce);'])

I was using your code as the framework for mine, which is why they are so similar. The main difference is that I manually inserted the column header names because my motion file does not include all of the coordinates that are in the model. Could this create an issue somehow?

Posted: **Tue Jul 21, 2015 12:21 pm**

Also:

Code: Select all

No appropriate method, property, or field equilibriateMuscles for class
org.opensim.modeling.Model.

Error in getNFL (line 68)
   myModel.equilibriateMuscles(s)

Is there an issue with installation if these are missing?

Posted: **Mon Aug 03, 2015 1:28 pm**

Your downcasting to a muscle that OpenSim doesn't support, hence the "_Deprecated". You will need to change the muscles to either Thelen or Millard types to use in scripting.

Posted: **Wed Aug 05, 2015 11:02 am**

Great thank you! That would explain why the two Thelen muscles in the model are returning normalized fiber lengths but the deprecated ones are not.

simtk.org

MATLAB returning zero fiber lengths

MATLAB returning zero fiber lengths

Re: MATLAB returning zero fiber lengths

Re: MATLAB returning zero fiber lengths

Re: MATLAB returning zero fiber lengths

Re: MATLAB returning zero fiber lengths

Re: MATLAB returning zero fiber lengths