Hi Peter
Using the MinMaxOOSolQP with a high penalty will most likely give you less
load sharing between the muscles that using a pure MinMaxSimplex with no
penalty . So if you change the criteria you may see a difference.
My guess is that the difference in muscle architecture (vastus creates
moment around the hip) and the load on the model, prevents these muscles to
be active at the same time in the time steps in question. The first thing to
investigate is to look at the moment reactions in the hip, if you run the
model with no muscles you will find a folder named JointMuscles here you can
read out the joint moments of the hip. Carefull investigation of the moment
balance around the hip and the knee may reveal why these muscles are not
active at the same time.
If you are using the MinMaxSimplex solver it ensures that there are no other
way the muscles can be recruited that will give a lower activity, this is
the optimal solution. If the VM,Vi, VL, RF are not all active when they
should be i would look for an explantion in the moment balance of the hip
joints.
You can also try to shift to the simple muscle model, just to rule that it
is the muscle properties which are playing a trick.
About the Eshuis knee: You can upload it but i can not garantee that anyone
will have time to solve the issues, as i recall it there was a reason why it
is not part of the repository, but maybee 
I hope this helps you move on, otherwise please write again.
Best regards
Søren
From: anyscript@yahoogroups.com [mailto:anyscript@yahoogroups.com] On Behalf
Of Peter Worsley
Sent: 29 January 2008 10:49
To: anyscript@yahoogroups.com
Subject: RE: [AnyScript] muscle modelling
Hio Soren,
thank you for your repsonse. I was using the MinMaxOOSolQP recruitement
solver with a QP penalty of 1000. In order to reduce force output of over
active muscles I changed the MusPar file by reducing the max ouptut at
optimial fibre length. I have also changed some anyvia points and insertion
points of muscles to get better results compared to literature.
My problem is this, in the human body any number of muscles will be
recruited for one given movement, but in the model it seems only one muscle
is designated for a given action (knee extension). Surely if the force was
shared across muscle groups (VM, VI, VL, RF) this would lead to a decrease
in activity per muscle.
In terms of the foot model I think I have realised the problem, the foot is
currently modelled as a universal joint allowing rotations about 2
perpendicular axes. But the current marker set (maleolus, calcaneus, 2nd
metatarsal) only controls movement in the sagittal plane, and the foot can
flip in the frontal plane.I am now working on adding extra points to the
foot markers to stop this.
One further question, John Rasmussen was kind enough to pass on the Eshuis
knee for me. But the model was written in version 2 script and it seems to
only work with closed chain activities (bike, sit-stand). I have got the
version changes from the anybody website but i am struggling to get the
model working. Could I post the file up on the forum for others to try?
Many thanks for your continued support.
peter
AnyBody Support <support@anybodytech <mailto:support%40anybodytech.com>
.com> wrote: Hi Peter
I will try answering your question, but I am afraid I also have some
questions for you
When you write that you have changed MVC in the over active muscles and that
this helped a bit, I do not understand how you did this in the model ?
You are right about the time steps these are independent so the previous one
do not affect the next one.
How did you specify a mean squared activity objective? Did you do this by
using the OOSolQP solver and a high penalty?
If the Rectus femoris and the vastus muscle do not work together in a
specific timestep it can be related to the moment balance around the hip
joint, since the rectus muscles in contrast to the vastus muscles inserts on
the hip. You are writing that you have some problems with foot, if the
force from the force plate is not positioned correctly on the foot this may
also affect the moment balance around the hip and the rectus fem activity,
but this is just a guess.
I hope this helped clarifying some of your questions.
Best regards
Søren
From: anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com
[mailto:anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com] On
Behalf
Of peter.worsley
Sent: 24 January 2008 13:02
To: anyscript@yahoogrou <mailto:anyscript%40yahoogroups.com> ps.com
Subject: [AnyScript] muscle modelling
Hi everyone,
i am currently using anybody to model gait in normal subjects. The
model is producing some very high axial loads in the knee joint due to
some large muscle force productions. Current marker positions in the
ankle are also creating problems (foot is flipping). I have changed
MVC in the over active muscle groups and this has helped with the
problem, but there still seems to be a pattern where force cannot be
shared amoung a muscle group (eg quadriceps). I know the time steps
don’t take previous contractions into consideration but I can’t
understand why the least mean squared activty model does not share the
muscle force across rec fem, vas med/lat/int during a given time frame.
Has anyone else been experiencing this porblem
many thanks
Peter
[Non-text portions of this message have been removed]
Users of the AnyBody Modeling System help each other create biomechanical
models in the AnyScript language.
Yahoo! Groups Links
Sent from Yahoo! - a smarter inbox.
[Non-text portions of this message have been removed]
[Non-text portions of this message have been removed]