Foward Dynamics

[Nicolas Schwenck]

Hej,

I am trying to use the forward simulation tool to simulate a motion driven by my own muscle excitation patterns. I am wondering if this is possible or if I have to record a motion first and then do all the steps that lead me to the CMC result?

I understand not quite right the meaning of the "states input" file. In the UsersGuide is said, that it is needed to compute the initial states, but do I need an initial state for every timestep?

[Ayman Habib]

Hi,

This is definitely possible (run forward simulation with your own excitations). For forward simulation you only need initial-states and excitations for the muscles if your model have some (regardless of where they come from).

If your motion requires external forces (e.g. ground reactions) however, you'll have to record these somehow or provide your own which would be "very hard" unless you record them.

The use of the inital states file is best explained by the tooltip of the corresponding field in the GUI. Only one is needed per simulation to initialize the system. From there on, OpenSim integrates the system to advance it in time.

Hope this helps,
-Ayman

[Nicolas Schwenck]

Ok, it sounds nice. I have tried to do a forward simulation with the example gait model 2354 and the original setup file. Before that I have done all the steps incl. CMC. Then I modified the initial states file in that way, that I deleted all the data except the first 5 timesteps. Unfortunately the Forward Dynamics tool did not complete and errors: "Forward integration did not complete. Keep partial results?". If I keep the partial result, then I see nothing.

Is there a more detailed information how the forward integration here works and what I have to provide in the states file?

yours
nicolas