Controls for forward dynamics generated by static opt.

[Yannik Schröder]

Hi,
i am currently trying to perform some static optimization calculations on an updated version of the wrist model i am working on at the moment. I generated a simple finger movement motion in a text editor which performs an extension of the index finger in the MCP joint:

Code: Select all

time 2_MCP_flex pro_sup
0	30	-28
0.1	25	-28
0.2	20	-28
0.3	15	-28
0.4	10	-28
0.5	5	-28
0.6	0	-28

Now i tried to calculate the muscle forces needed to perform the movement via static optimization. But first i locked all joints but the ones of the index finger and the wrist. After the SO calculation i get a new motion file calles "states". What exactly is this motion? And how does it differ from the original motion i am trying to get the muscle forces for?

Then, to validate the results, i tried to perform a forward dynamics simulation with the control.xml file generated by the static optimization as an input. I expected the model to perform the extension motion i just calculated the control set for, but this is not the case. The motion differs much from the original motion. Why is that? Currently i used the following Workflow:
Generate motion file -> Calculate the activations (=controls.xml) of the muscles via SO for the motion -> Use the generated controls.xml to perform a FD calc. and compare the outcome with the original motion.

Can't i use forward dynamics to validate the calculated control set? If not, how can i check whether my results are plausible? Does a low performance value automatically mean my results are good?

And an additional question: the more joints i unlock the more "muscles are to weak" warnings appear. But i got my muscle values out of the literature so they should be, more or less, correct. What could be the origin of these errors? Wrong muscle paths and resulting wrong lever arms?

Thanks in advance!

Best regards,
Yannik