Hi Rudi,
thanks for your reply. It works well for 0.05 s. However, I have tried to increase the duration of the simulation, and after 0.05, I had a large deviation. My problem is that 0.05s is too short to make forward dynamics during walking trial.
Finally, the solution is to run the FD based on torque only during 0.05s, and use the final state as initial state and run FD on the next simulation time.
I thought that it was possible to run Foward Dynamics for a stance phase of walking directly, and it seems that it's not possible.
Thanks for your help,
Pauline