simtk.org

Hi Guys,

I have an issue with the computeStateVariableDerivatives command. Both in C++ and Matlab, and only for certain states, the handle returns NaNs or -1.#IND. It will return these values for ALL acceleration values within the resulting state derivatives variable.

Therefore I am wondering when this function returns NaNs, is it an issue with too small/ big nrs, or something else entirely?

Hope someone can give me some details on this function!

Regards,

Lars Boers

Hi Lars,

Are you initializing the system and setting the state variables prior to calling computeStateVariableDerivatives?

I do something like this in Matlab:

% Import the OpenSim modeling classes
import org.opensim.modeling.*

% Read in the osim model
osimModel = Model('aModel.osim');

% Initialize the model
osimState = osimModel.initSystem();

% Get the number of states;
nStates = osimModel.getNumStateVariables();

for i = 0:nStates-1
osimState.updY().set(i,states(i+1,1)); % 'i' is the index of a particular state variable & states is a vector of
% state variable values which you want to set
end

% Calculate derivatives
osimModel.computeStateVariableDerivatives(osimState);

Hope that helps!

Nicos