simtk.org

Hi Kat,

Thanks for a great plug-in! In your webinar you mentioned that activations could be weighted by changing their constant in the synergy matrix. You said changing a muscle's constant from 1 to 0.1 for example would essentially 'penalize' activation of that muscle. Can you describe in a little more detail what exactly occurs in OpenSim when an activation is 'penalized'? How is the way muscle activations are being estimated change? Also, how does changing the synergy matrix constant scale the level of penalization? If that constant is changed from 1->0.5, is the muscles activation penalized by a factor of 2 compared to a factor of 10 if it was changed from 1->0.1?

Thanks for the help!

Just like static optimization, synergy optimization minimizes the total sum of activations squared. However, while in static optimization all actuators are weighted equally, in synergy optimization you can provide a weighting constant in front of each actuator or group of actuators. So you are minimizing the sum of (c*a)^2 where c is the weighting coefficient in the matrix and a is the activation. All of the weights are relative to one another (very similar to scaling or inverse kinematics where weights are used for markers) - so values of 0.1 and 1 would be the same as setting up a different matrix with values of 1 and 10. Hope that helps.