Summary: AnyReacForce

Ok, thank you for your patience, i think, now i got it.
In summary i understood the following:

There are two conditions that have to be fulfilled:

#1 the desired posture (defined in mannequin.any)
#2 the equilibrium

AnyReacForce generates moments and is responsible for #1.
The muscles care for #2.
And of course there is interaction between #1 and #2, but the
calculated solution fulfills the objective function. (Different
moments would lead to higher muscle activation.)

My thinking was, that the muscles care for both, the desired posture
AND the equilibrium.

Please correct me, if i am still wrong.

Thanks, Thomas

Hi Thomas

You are right about the two conditions that has to be fulfilled; the posture
and force equilibrium.

But it is not correct that the AnyReacForce is responsible for the posture.
The AnyReacForce object can only be used for providing reaction forces, in a
way that does not provide any motion. This is in contrast to the drivers in
the model which will always supply a motion and can supply forces if this is
switched in, by the Reaction.Type setting

AnyBody is running inverse dynamics so the motion is input to the model; the
motion is not influenced by the forces. In most of the repository models the
motion is provide as a combination of mannequin joint angles, (which is then
used by drivers) and connections to the environment.

The force equilibrium is taken care of by the muscles and the reaction
forces in the model. The reactions come as joint reactions or as
AnyReacForces. The muscle recruitment algorithm will try to lower the
highest activated muscle; in this process the reaction forces will be
determined automatically. The reaction forces has no cost in the objective
function so the algorithm will always try to use the available reactions in
order to lower the muscle activations as much as possible.

So the solution is in principle a two step procedure, first the kinematics
is solved then the kinetics is resolved.

It is actually also possible to add an extra optimization loop around the
model, which will allow the forces to influence the motion, but this
requires a user defined objective function, this kind of analysis we call
inverse-inverse analysis. As an example of this please see
http://www.anybody.aau.dk/pdf/inverse.pdf where the motion of the ankle
joint in a bicycle model has been optimized using this approach

I hope this made things clearer, otherwise please write again.

Best regards

Søren, AnyBody Support


From: anyscript@yahoogroups.com [mailto:anyscript@yahoogroups.com] On Behalf
Of zandfub1
Sent: 15 October 2008 15:29
To: anyscript@yahoogroups.com
Subject: [AnyScript] Summary: AnyReacForce

Ok, thank you for your patience, i think, now i got it.
In summary i understood the following:

There are two conditions that have to be fulfilled:

#1 the desired posture (defined in mannequin.any)
#2 the equilibrium

AnyReacForce generates moments and is responsible for #1.
The muscles care for #2.
And of course there is interaction between #1 and #2, but the
calculated solution fulfills the objective function. (Different
moments would lead to higher muscle activation.)

My thinking was, that the muscles care for both, the desired posture
AND the equilibrium.

Please correct me, if i am still wrong.

Thanks, Thomas

[Non-text portions of this message have been removed]