Hi Prabhav
I have tried your model again and done some updates.
First of all the Direction variables in the bottom of the file was not
correct it should be Direction ={2,0,1}; because the z direction is the
normal direction.
Secondly the UserDefinedLimitLow of the conditional muscles should be
lowered to ensure that there is always a force there, I used -2…
I also at some point changed the gravity slightly to increase the small
terms … Gravity = {0.0001, 0.001,-9.81};
Also, there was no contact element in the big toe.
Finally there was an AnyReacForce between the first toe segment and the
CenterofPressure segment for the toes on all six dof., this should not be
there.
If it is there with all six dof. the entire load will go directly to the
first toe, with out using the artificial muscles.
AnyReacForce ForcePlateToesFootRContactForce={
// AnyKinLinear LinT={
// //very very important this ensurs that the reaction is between
artifical segment and the foot
// //and not between foot and globalref if Ref had been equal -1 which
is default
// Ref=0;
// AnySeg &ref1=…CenterOfPressureToes;
// AnySeg &ref2=…HumanModel.FootSegmentsRight.Toes;
// };
AnyKinRotational RotT= {
AnySeg &ref1=..CenterOfPressureToes;
AnySeg &ref2=...HumanModel.FootSegmentsRight.Toes;
Type=RotVector;
};
};
I tried to remove it entirely but then it can not recruit so I kept the
moment which is not a nice solution.
The problem is I suspect at least that there is no way for the pushing
muscles to balance the load since it for some reason requires a pulling
force. Looking at the modelview it looks like the applied force is place
more medial than the contact point on the big toe this makes it difficult
for the muscles to carry the load, when they can only push.
To make this setup work you have to ensure somehow that the center of
pressure element can be carried by the pushing muscles, and if the applied
force is "outside " the area the contact points define then the pushing
muscles will not be enough to carry the load.
At the moment the only idea I have is to introduce more contact element on
the toes that will create a bigger “contact area” this would make it easier
to balance the center of pressure element.
I have also tried out other solutions, but right now this is the best idea I
have for this.
Hope it made sense and help you move on.
Best regards
Søren, AnyBody Support
-----Original Message-----
From: anyscript@yahoogroups.com [mailto:anyscript@yahoogroups.com] On Behalf
Of Prabhav Saraswat
Sent: 09 December 2008 01:14
To: anyscript@yahoogroups.com
Subject: Re: [AnyScript] Problems with Foot Model.
Hi Soren,
About changing the strength object, I changed it becasue it was easier to
define “UserdefinedLimitLow & High” relative to glabal origin instead of
Target object. My understanding was that Strength obeject is used to
calcualet min/max distance at which reaction will be turned on. I changed it
back to what you suggested and I am still getting the same error.
CenterOfPressure should be at the right position when you run
inversedynamics in OptSudy. Can you please explain what you mean by “normal
direction of segment”?
Thanks
Prabhav
On Mon, Dec 8, 2008 at 5:29 AM, AnyScript Support <
anyscriptsup@anybodytech.com> wrote:
> Hi Prabhav
>
> I tried to replicate this error but I get a warning saying that the model
> is
> kinetically over determined and therefore no muscles are loaded.
>
> Looking into the drivers file “MADrivers.any” I found that all drivers
have
> reactions switched on, so they will carry the load instead of the muscles.
>
> You have to set some of these to off, my guess would be that you would
like
> to carry the femur entirely and then have all the other drivers not to
> carry
> any load? If you have recruitment problems you can also keep some of them
> while debugging the model, but right now you will not see any muscle load
> at
> all.
>
> I also looked in the environment2.any file and found some issues,
> concerning
> the new conditional elements you have created.
>
> First of all if you insert a reference frame in the BaseObject which is
the
> CenterofPressureSegment for the toes it draws this reference frame I a
> point
> close to the knee, this segment should be located somewhere below the toes
> for this setup to work. Please see the code below on how the visible
> reference frame could be added.
>
> AnyFolder ConditionalContactToe2 = {
> BaseObject ={AnyDrawRefFrame drw={};};
> AnyRefFrame &BaseObject=.CenterOfPressureToes; //Object which delivers
> the forces
> AnyRefFrame &TargetObject=
> …HumanModel.FootSegmentsRight.Toe2;//Location where the force is applied
> AnyRefFrame &StrengthObject=TargetObject; //Node used for strength
> measurement, occasionally this is different from the target object
>
> In the lines above I also changed the line about the strength object to be
> AnyRefFrame &StrengthObject=TargetObject;
> Please remember to change this in your model, for this kind of model it
> should be the same as the Targetobject.
>
> Once you have the CenterOfPressureToes segment in the right location you
> need to make sure you know what is the normal direction of this segment?,
> If
> the x direction is the normal then the Direction vector you have specified
> in the conditional contact element is ok (Direction={0,1,2};), otherwise
it
> has to be changed, to for example Direction={1,0,2}; if the y direction is
> the normal.
>
> Best regards
> Søren, AnyBody Support
>
> -----Original Message-----
> From: anyscript@yahoogroups.com <anyscript%40yahoogroups.com> [mailto:
> anyscript@yahoogroups.com <anyscript%40yahoogroups.com>] On Behalf
> Of Prabhav Saraswat
> Sent: 07 December 2008 02:29
> To: anyscript@yahoogroups.com <anyscript%40yahoogroups.com>
> Subject: Re: [AnyScript] Problems with Foot Model.
>
> Hi Soren,
> Thanks for your quick response. I tried to implement the contact element
> for
> 4 toes. But I am getting this error when I run inverse dynamics-
>
> ERROR : Unexpected error in system kernel :
OOSmatrix::operator()(int,int)
> : Index of vector or matrix is out of range.
>
> What could i have done wrong to get this strange error? I have uploaded
the
> model with changes here-
>
> http://www.yousendit.com/download/TTZuYURBcG96NEozZUE9PQ
>
> Thanks
> Prabhav
>
> On Tue, Dec 2, 2008 at 7:44 AM, AnyScript Support <
> anyscriptsup@anybodytech.com <anyscriptsup%40anybodytech.com>> wrote:
>
> > Hi Prabhav
> >
> > Some more comments on the load distribution between the toes…
> >
> > The reason why the setup with the reaction forces does not work for the
> > distribution of the forces between toes is that the reaction forces has
> no
> > cost in the objecttive function. So if you have dof’s which can be
> carried
> > by several reactions you do not know which of these that will carry the
> > load. In one timestep it can be one of them in another step the other
> one.
> >
> > The way to fix this is to introduce artificial muscles instead of the
> > reaction forces to the dummy segment. If you look in the
> > jointanddrivers.any
> > file of the pedaldemoconditional example you will find an example of a
> > contact element, that can be used for creating contact between objects.
> If
> > you introduce one of these elements acting from the dummy segment to the
> > each of the toes and remove the lin and rot reactions i think it will
> work.
> > This contact element has a normal force and friction forces linked to
the
> > normal force, it will not allow “sticky” forces and moments are not
> > carried,
> > but if you have five set of these it might be enough to carry the
moments
> > of
> > the dummy segment.
> >
> > The contact element is made up using a bunch of general muscles so it
> will
> > add a cost to the objective function. You can play around with the
> Strength
> > setting, if it is very high it will act more and more like a true
> reaction,
> > whereas the lower it is the more it will try to smooth out the load
> between
> > the toes.
> >
> > Hope it made sense 
> >
> > Best regards
> > Søren, AnyBody Support
> >
> > _____
> >
> > From: anyscript@yahoogroups.com <anyscript%40yahoogroups.com><anyscript%
> 40yahoogroups.com> [mailto:
> > anyscript@yahoogroups.com <anyscript%40yahoogroups.com> <anyscript%
> 40yahoogroups.com>] On Behalf
> > Of AnyScript Support
> > Sent: 02 December 2008 11:35
> > To: anyscript@yahoogroups.com <anyscript%40yahoogroups.com> <anyscript%
> 40yahoogroups.com>
>
> > Subject: RE: [AnyScript] Problems with Foot Model.
> >
> >
> > Hi Prabhav
> >
> > This looks like it could be a very nice model
> >
> > I have looked through some of it and found some issues…
> >
> > When you load the model you get a warning that there are redundant
> > constraints, this should not be ignored since it may have an impact on
> the
> > results.
> >
> > Looking through the model I found that the reason for this error is the
> > load
> > applied to the toes, here you have one segment with the applied load and
> > then you create six reactions form this segment to each of the toes,
this
> > will not work. The reason is that this setup creates constraints between
> > the
> > toes, because they all have reactions to the same dummy segment, so
> forces
> > can go from one toe through the dummy segment to the other toes. The way
> to
> > resolve this is to divide the load into five contributions and apply it
> > using five segments, just as you are doing with the other loads.
> >
> > I tried to remove all the toe reactions except the one acting on toe1
but
> > this leads to recruitment problems, looking at the muscle configuration
> > this
> > is not so surprising since it does not look like that the individual
toes
> > have muscles capable of both flexion and extension.
> >
> > I think you should try following these steps.
> >
> > 1 Switch to using simple muscles
> >
> > 2 add temporary joint reaction until the model runs inverse analysis
> >
> > 3 replace these temporary reactions with real muscles one by one until
> they
> > are all removed and the model is running inverse analysis
> >
> > 4 Switch back to the mus3e configuration, run the kinematic analysis and
> > look at the strength variations of the muscles, I tried to do so and it
> > looks like that many of them becomes very weak …
> >
> > 5 If the activation remains to high then try to add temporary reactions
> to
> > the joints one by one, if the activations drops it may tell you which
dof
> > is
> > the bottleneck and you can focus on addressing the muscle spanning this
> > dof.
> >
> > Hope this helps you move on
> >
> > Best regards
> >
> > Søren, AnyBody Support
> >
> > _____
> >
> > From: anyscript@yahoogrou
> <mailto:anyscript%40yahoogroups.com <anyscript%2540yahoogroups.com>
> <anyscript%2540yahoogroups.com>>
> > ps.com
> > [mailto:anyscript@yahoogrou
> <mailto:anyscript%40yahoogroups.com <anyscript%2540yahoogroups.com>
> <anyscript%2540yahoogroups.com>>
> > ps.com] On
> > Behalf
> > Of Prabhav Saraswat
> > Sent: 01 December 2008 23:44
> > To: anyscript@yahoogrou
> <mailto:anyscript%40yahoogroups.com <anyscript%2540yahoogroups.com>
> <anyscript%2540yahoogroups.com>>
>
> > ps.com
> >
> > Subject: [AnyScript] Problems with Foot Model.
> >
> > Hi all,
> > I have been working on 3-segment Foot model and have a few problems with
> > the
> > model outcomes. I have uploaded the model here, it was too large to
> upload
> > on the group website-
> >
> > http://rcpt.
> > <http://rcpt.
> > <http://rcpt.yousendit.com/630258784/b755d231e9e8f8ec32b16d42af8f6406>
> > yousendit.com/630258784/b755d231e9e8f8ec32b16d42af8f6406>
> > yousendit.com/630258784/b755d231e9e8f8ec32b16d42af8f6406
> >
> > Model is driven by normal gait analyis data. It has spherical joint for
> > ankle & midfoot and revolute joint for toes. Ground reaction force is
> > divided into 3 sections and applied on each segment.
> >
> > These are the major issues with model-
> > - Some of the muscles show very high activity level.
> > - Few muscles (Tibialis Posterior, Peroneus Tertius, EHL, FDL) are not
> > getting recruited. I checked the moment arms of these mucles and they
> look
> > fine as compared to cadaver studies.
> > - Some muscle activation (Gastroc & Soleus) is delayed as compared to
> > normally expected activation pattern for gait.
> >
> > Any suggesions on what might be wrong with the model will be helpful.
> >
> > Thanks
> > Prabhav
> >
> > [Non-text portions of this message have been removed]
> >
> > [Non-text portions of this message have been removed]
> >
> > [Non-text portions of this message have been removed]
> >
> >
> >
>
> [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
>
>
>
[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