FDK with MinMax criterion

Hi,

I’m trying to use the FDK algorithm with the min/max muscle criterion for a gait trial. Compared to the polynomial criterion, the min/max seems to take more time to find a solution and sometimes fails to find one, while the polynomial criterion is faster and more robust. Is it just an impression ? Is there a rational behind it ?

I think that the min-max criterion has the tendency to estimate too low activities at the begining of the motion, which is not so great for the FDK convergence. I tried to add a lower bound of 1% on muscle activity in the Inverse dynamics settings (i.e. InverseDynamics.Criterion.LowerBoundOnOff = On; InverseDynamics.Criterion.LowerBound = 0.01). It does not seems to help much.

Do you have any suggestions that would help the FDK to converge ?

Thank you.

Pierre