Hi Eugene
I yes i think you should increase the strengh of the muscle group
with the same amount unless you know that the subject have
particular strong legs compared to the rest of the body.
There can be multiple reason why your model do not follow the
markers exactly. If they should follow exactly there are a number of
things which should be correct:
1 the sizes of the bones
2 the locations of the markers on the segments
3 the kinematics of the joints in the model should be the same as in
the subject, here the primary error source could be the knee joint
since this is modeled as a hinge joint. The ankle joint is similary
made as universal joint. So these joint types are simplified and
this may lead to errors between the markers.
4 if there are any skind artifacts on the markers this may lead to
errors as well, since the markers on the humanmodel are rigidly
connected.
So to improve the model these are the items to look into, and here
the first two are the most obvious to start out with.
The Leg model has in total 18 dof
6 rigid body dof
2x3 in the hips
2x1 in the knee
2x2 in the ankle
This means that you can not drive more dof than these 18 dof, if you
attempt to do so it will make the model overconstrainted and it will
not work kinematically.
This is the exaplanation why only certain dof from each marker can
be used, the number os drivers must add up to 18 no more, no less.
If for example we made use of all makers of the knee instead of the
single one we have now we would need to free up some other dof in
the model.
If the knee is the most important joint in your work it migth be
that it will be posible to replace two of the drivers in the heel
with two driver of the knee. This might make the knee better defined
but it will leave the heel less well defined.
It is difficult to say the motion is good enough, but you may do a
sensitivity analysis on your results using different kinematic
drivers to see how the kinematics influence the forces in the knee.
I hope this helps you move on, otherwise please write again.
Best regards
Søren, AnyBody Support
— In anyscript@yahoogroups.com, “eugeneyang40” <eugeneyang40@…>
wrote:
>
> Hi Soren
>
> I have some question which I am not sure how to answer and need
your
> advise.
>
> For my model, the StrenghtIndexLeg is 1.5, and
> SpecificMuscleTensions is currently at 90. You suggest that I
should
> increan the SpecificMuscleTension by another 45?
>
> How come I cannot make my model follow the markers exactly? since
my
> model does not follow all the marker exactly, how can I be sure
that
> my calculation for Knee Reaction force is reasonable. The model is
> only constraint to certain axis while the rest are free, why is
that
> so? And how come I must drive the model in those constraint?
>
> Thank you.
>
> Regards
> Eugene
>
> — In anyscript@yahoogroups.com, “AnyBody Support” <support@>
> wrote:
> >
> > Hi Eugene
> >
> > It is difficult to say what the exact number should be for the
> > IndexLeg, if your subject was an athlet. If you increase the
> > strenght of the legs by a cetain pertcentage i think you should
> > increase the SpecificMuscleTensions by the same amount. As i
> recall
> > the numbers from your model the activty was around 1.6 so
> increasing
> > by 60% would be a guess. As long that you scale the individual
> body
> > parts by the same amount the only difference it will make on
the
> > results is that it will change the activity level. The shape of
> the
> > curves will remain the same.
> >
> > The best way to make the model and the subject fit in terms of
> > strength, would be to make some strength test on the subject and
> the
> > make the model replicate these experiments. Then optimizing the
> > strength paramters of the muscles to make the best match, but
this
> > solution is obviously not straight forward to do.
> >
> > If you would like to model different persons you will have to
> > customize it for each person, that is correct. This is done by
> > adjusting the segment lengths in the AnyMan.any file. The
segment
> > lengths are in general the length from joint to joint. The
pelvis
> > width is the distances between the hipjoints. Unfortunately i
can
> > not provide much help on how to do the measurement, i guess you
> > would use a simpel tape measure. There is a new book by Serge
van
> > Sint Jan, without having seen it, i think it migth be usefull
for
> > this purpose. The title is:
> >
> > Color atlas of skeletal landmark definitions : guidelines for
> > reproducible manual and virtual palpation
> >
> > here is a link
> > http://www.us.elsevierhealth.com/product.jsp?isbn=9780443103155
> >
> >
> > Best regards
> > Søren, AnyBody Support
> >
> >
> > — In anyscript@yahoogroups.com, “eugeneyang40” <eugeneyang40@>
> > wrote:
> > >
> > > Hi Soren,
> > >
> > > Thanks for the help. It seem to be moving fine. Regarding the
> > > StrengthParameters, how do i know that the Strength indexLeg
is
> > > correct and what value should i set for
> SpecificMuscleTensionSpine
> > > and SpecificMuscleTensionShoulderArm.
> > >
> > > Actually I wanted to use this model for different subjects,
but
> i
> > > guess i must customize each model for different subject right?
> > This
> > > movement of this model was actually done by me. If I wish to
do
> a
> > > female athlete, I must amend the Segment Length in AnyMan.any
or
> > use
> > > AnyArne.any right? I just wish to check for the ThinghLength,
> > > ShankLength and FootLength and PelvisWidth, how should the
> > > measurement be taken? So that i can get a more accurate
> > SegmentLength.
> > >
> > > Thank you.
> > >
> > > Best Regards
> > >
> > > Eugene
> > >
> > >
> > >
> > > — In anyscript@yahoogroups.com, “AnyBody Support” <support@>
> > > wrote:
> > > >
> > > > Hi Eugene
> > > >
> > > > I have tried to improve the kinematics of your model by
making
> > > > changes to the following files
> > > >
> > > > JointAndDrivers.any
> > > > Here the drivers setup now more like the original Gait3D
> model.
> > > > I had to free up drivers on the pelvis in order to be able
to
> > > apply
> > > > drivers on the knees.
> > > > Added a reaction forces on the Eversion/inversion of the
> ankles
> > > > sometimes the msucles have a problem carrying this dof.
> > > >
> > > >
> > > > KitVaughanMakers.any
> > > > I have move the markers of the Asis and the sacrum a bit
upward
> > > >
> > > > AnyMan.any
> > > > The segment lenghts have been shortened a bit
> > > >
> > > > MyLanding2L.Main.any
> > > > Changed the scaling to be LengthMassFat
> > > > Outcommented the KinematicTol so that the default one is
being
> > used.
> > > >
> > > > I will upload the model shortly
> > > >
> > > > After having seen the motion i think the problem of the
> > illiopsoas
> > > > muscles is the that the muscle is wrapping on the “end-egde”
> of
> > the
> > > > cylinder. This can cause problems for the algorithm and may
> > require
> > > > more points like you have noticed. A better solution is to
> make
> > the
> > > > cylinder a bit longer so that the muscle wrapping on the
> > cylinder
> > > > edge is avoided.
> > > >
> > > > This can be done in the file
> > > > Leg.Root.any in the folder named “pecent_ossis_pubis” by
> > > multiplying
> > > > the "Length " property by a factor of two. If you are using
> the
> > > > other muscle model this should also be done in that root
file.
> > > >
> > > >
> > > > For the Leg model only the index variable is relevant, by
> > adjusting
> > > > this you can increase the strength of the leg muscles. If
your
> > > > subject was an athlete the index should be increased. The
> reason
> > > why
> > > > the leg is different from the other body parts is that in
the
> > legs
> > > > we have the strengh given directly as a number whereas in
the
> > other
> > > > body parts we use PSCA * Stress.
> > > >
> > > > I hope this helsp you move on, it is a really nice model!
> > > >
> > > > Best regards
> > > > Søren , AnyBody support
> > > >
> > > >
> > > > — In anyscript@yahoogroups.com, “eugeneyang40”
> <eugeneyang40@>
> > > > wrote:
> > > > >
> > > > > Hi Soren,
> > > > >
> > > > > I had attached a new folder name “Landing Gait” of my
single
> > leg
> > > > > landing model. I did not use the same constraint as in
> Gait3D.
> > I
> > > > > still have problem as the knee does not go with the
markers.
> I
> > > try
> > > > > the amendment you advised but it does not seem to work.
For
> > the
> > > > > iliopsoas muscle,I got to increase the SPLine.StringMesh
to
> > about
> > > > 130
> > > > > before my model can work. I not sure what went wrong with
my
> > > > model. I
> > > > > also included AnyReactionForce for the Knee and ankle to
> > measure
> > > > the
> > > > > joint force.
> > > > >
> > > > > For the muscle strength, what will be an accurate value to
> be
> > > set?
> > > > I
> > > > > understand from the tutorial that it is set at 90 N/cm2
and
> > the
> > > > index
> > > > > at 1. I not sure does this differ for different type
> consider
> > my
> > > > > experiment subject is an athlete.
> > > > >
> > > > > Thank you.
> > > > >
> > > > > Regards
> > > > > Eugene Yang
> > > > >
> > > > > — In anyscript@yahoogroups.com, “AnyBody Support”
> <support@>
> > > > > wrote:
> > > > > >
> > > > > > Hi Eugene
> > > > > >
> > > > > > Here are some answers to your questions:
> > > > > >
> > > > > > 1 If you have not done modifications to the original
> > > > > > JointAndDriver.any file the knee is driven in the X
> > direction
> > > of
> > > > > the
> > > > > > thigh. This direction is posterior/anterior on the
thigh.
> If
> > > the
> > > > > > knee markers do not follow the markers, this can be be
due
> > to
> > > > > > kinematics other places in the model. If for example
the
> > legs
> > > > are
> > > > > > too long the knee will not follow the knee marker, it
will
> > only
> > > > do
> > > > > > this for the x dierection, the y and z directions
> > are “free”.
> > > > These
> > > > > > are determined by the other markers, the human joint and
> the
> > > > > > segments lengths. In other words they are indirectly
> driven.
> > > > > >
> > > > > > So i think you should closely evaluate if the segments
> > lenghts
> > > > are
> > > > > > ok, and if the marker locations used in your data
> > corresponds
> > > to
> > > > > the
> > > > > > actual location of the markers in the model. A simple
> check
> > is
> > > > to
> > > > > > find two markers on the same segment and check the
> distance
> > > > between
> > > > > > these markers in the recorded mocap data and on the
model,
> > > these
> > > > > > should match resonable well.
> > > > > >
> > > > > > Another issue is the starting position of the model, due
> to
> > the
> > > > way
> > > > > > it is driven with closed kinematic chains the staring
> > positions
> > > > of
> > > > > > the segments are important. These are controlled through
> the
> > > > > > mannequin.any file. The angles set in
the “mannequin.any”
> > file
> > > > are
> > > > > > only important for the first kinematic step
> > > > > >
> > > > > > 2: Please see the respond just posted on group in the
> > previous
> > > > > > message
> > > > > >
> > > > > > 3: The different scaling laws may change the kinematics
of
> > the
> > > > > > model. If the model is difficult to assemble
kinematically
> a
> > > > change
> > > > > > in the scaling law may cause it to fail. Again the
remedy
> is
> > to
> > > > > look
> > > > > > at segment lengths, marker locations, and initial
> positions,
> > > > > >
> > > > > > I hope this helps you move on.
> > > > > >
> > > > > > Otherwise please write gain.
> > > > > >
> > > > > > Best regards
> > > > > > Søren, AnyBody support
> > > > > >
> > > > > > — In anyscript@yahoogroups.com, “eugene yang”
> > <eugeneyang40@>
> > > > > > wrote:
> > > > > > >
> > > > > > > Hi Christian,
> > > > > > >
> > > > > > > Thank for your previous suggestion. Correct the gray
> > markers
> > > > are
> > > > > > what i
> > > > > > > obtain using MOCAP and the blue are those specify in
Kt
> > Van.
> > > > > > Markers. I redo
> > > > > > > the model again with only the right leg and the then
> both
> > > > leg. I
> > > > > > manage to
> > > > > > > run the model at Kinematic Tol of 0.005. But I face a
> few
> > > > > > problems. Both of
> > > > > > > them i used uniform scaling.
> > > > > > >
> > > > > > > For the Single Leg:
> > > > > > >
> > > > > > > 1) The Knee movement does not follow the knee markers
> > > > > > >
> > > > > > > For the two leg model:
> > > > > > >
> > > > > > > 1) The Knee does not follow the markers after I set
the
> > > > drivers.
> > > > > > >
> > > > > > > 2) I was unable to run the model with muslce. It show
> the
> > > > error
> > > > > > msg:
> > > > > > > Iliopsoas : Unexpected exception in the library
> > OOSol :
> > > > OOSol
> > > > > > exception
> > > > > > > : General factorization exception : Negative or zero
> pivot.
> > > > > > >
> > > > > > > 3) I can only run the no muscle model with uniform
> > scaling.
> > > > When
> > > > > i
> > > > > > applied
> > > > > > > the LengthMass Fat Scaling my kinematic error becomes
> very
> > > > large.
> > > > > > >
> > > > > > > Thanks
> > > > > > >
> > > > > > > Best Regards
> > > > > > > Eugene Yang
> > > > > > >
> > > > > > > >From: “johnzengwu” <jwu@>
> > > > > > > >Reply-To: anyscript@yahoogroups.com
> > > > > > > >To: anyscript@yahoogroups.com
> > > > > > > >Subject: [AnyScript] Re: Introduction and above Error
> > > > tolerance
> > > > > > > >Date: Mon, 02 Jul 2007 19:22:43 -0000
> > > > > > > >
> > > > > > > >Christian:
> > > > > > > >
> > > > > > > >I have a question about the scaling. I understand
that
> > the
> > > > > > LengthMass
> > > > > > > >or LengthMassFat is more reasonable than the Uniform
> > > scaling.
> > > > > > But, I
> > > > > > > >got problem when switch from Uniform to
> assFatLengthMass
> > or
> > > > > > LengthM
> > > > > > > >scaling. I got error messages:
> > > > > > > >
> > > > > > > >C:\My_document\AnyBody\Repository.6
> > > \BRep\Aalborg\Leg3D\Mus.any
> > > > > > (203) :
> > > > > > > > Iliopsoas : Unexpected exception in the library
> > OOSol :
> > > > > > > > OOSol exception : General factorization
exception :
> > > > Negative
> > > > > > or
> > > > > > > >zero pivot
> > > > > > > >
> > > > > > > >It seems that muscle Iliopsoas has zero strength or
> > negative
> > > > > fiber
> > > > > > > >length. My model works OK in uniform scaling. I
cannot
> > > figure
> > > > > out
> > > > > > the
> > > > > > > >problem. I have tried your sample Gait3D to apply
> > different
> > > > > > scalings.
> > > > > > > >It also does not work by using LengthMass or
> > LengthMassFat
> > > > > > scaling.
> > > > > > > >
> > > > > > > >Regards,
> > > > > > > >
> > > > > > > >John
> > > > > > > >
> > > > > > > >
> > > > > > > >— In anyscript@yahoogroups.com, “AnyBody Support”
> > > > <support@>
> > > > > > wrote:
> > > > > > > > >
> > > > > > > > > Hi Eugene Yang
> > > > > > > > >
> > > > > > > > > Thanks for the model. I downloaded it and took a
> look.
> > It
> > > > > does
> > > > > > run but
> > > > > > > > > do not move because of the large kinematicTol you
> set
> > on.
> > > > If
> > > > > > you
> > > > > > > > > change the kinematicTol to 0.0005 you can run
> > > > > > SetInitialConditions and
> > > > > > > > > you can see which markers that does not overlap.
You
> > made
> > > > all
> > > > > > the
> > > > > > > > > measured markers gray right? and some you made
lines
> > > > between
> > > > > > the mocap
> > > > > > > > > and the Kit Vaughan markers. It could look like
you
> > have
> > > a
> > > > > > scaling
> > > > > > > > > problem, seems like your model i bigger then the
> mocap
> > > > > person.
> > > > > > I see
> > > > > > > > > you are using uniform scaling, this is not a good
> way
> > to
> > > > > scale,
> > > > > > > > > instead try to use
> > > > > > BRep\Aalborg\Scaling\ScalingLengthMassFat.any which
> > > > > > > > > scales the model to a height and a weight. Hope it
> > helps
> > > > you
> > > > > > to move
> > > > > > > > > one step further, else you are very welcome to
write
> > > again.
> > > > > > > > >
> > > > > > > > > Best regards
> > > > > > > > > Christian, AnyBody Support
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > — In anyscript@yahoogroups.com, “eugeneyang40”
> > > > > > <eugeneyang40@> wrote:
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > Hi Christian,
> > > > > > > > > >
> > > > > > > > > > Thanks for the advise. I have attached a copy of
> > > > the “Copy
> > > > > of
> > > > > > > > > > Gait3D” i edited in the file section. I still
have
> > > > problem
> > > > > > running
> > > > > > > > > > the kinematic analysis. I used the force plate
> data
> > as
> > > > of
> > > > > > Gait3D as
> > > > > > > > > > my forceplate was not working during the MOCAP.
> > > > > > > > > >
> > > > > > > > > > The single landing move is as follows: It from a
> > double
> > > > leg
> > > > > > squat
> > > > > > > > > > jump and then landing on the right leg. The
> landing
> > is
> > > > like
> > > > > > a single
> > > > > > > > > > leg squat and the left leg is kept to the side.
> The
> > > data
> > > > i
> > > > > > used was
> > > > > > > > > > when the right foot touches the ground till the
> > squat.
> > > > > > > > > >
> > > > > > > > > > I changed the scaling to AnyArne.any instead of
> > > > > AnyMan.any,
> > > > > > and
> > > > > > > > > > removed all the forceplate2 data. But I am still
> not
> > > > sure
> > > > > > how come
> > > > > > > > > > my error is still so large and i cannot run my
> > model.
> > > > The
> > > > > > axis i
> > > > > > > > > > used is similar to that of Gait3D for my
qualisys
> > data.
> > > > > When
> > > > > > i
> > > > > > > > > > include muscle to the model and run kinematics
> test
> > it
> > > > have
> > > > > > a lot of
> > > > > > > > > > error regarding the muscle length.
> > > > > > > > > >
> > > > > > > > > > My apology for my poor knowledge in Anybody.
> > > > > > > > > >
> > > > > > > > > > Best Regards,
> > > > > > > > > >
> > > > > > > > > > Eugene Yang
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > — In anyscript@yahoogroups.com, “AnyBody
> Support”
> > > > > > <support@>
> > > > > > > > > > wrote:
> > > > > > > > > > >
> > > > > > > > > > > Hi Eugene
> > > > > > > > > > >
> > > > > > > > > > > Welcome to the group.
> > > > > > > > > > > Your project sounds interesting. Well depends
on
> > how
> > > > you
> > > > > > track your
> > > > > > > > > > > markers and in what direction you track them.
> How
> > > does
> > > > it
> > > > > > look when
> > > > > > > > > > > you run the kinematic analysis? or you cannot
> run
> > it
> > > > at
> > > > > > all?
> > > > > > > > > > > you can add a “KinematicTol” in the study to
> allow
> > > > some
> > > > > > tolerance,
> > > > > > > > > > > then you can maybe see where the problem is
and
> > then
> > > > fix
> > > > > > it.
> > > > > > > > > > > Hope it helped a bit?
> > > > > > > > > > >
> > > > > > > > > > > Best regards
> > > > > > > > > > > Christian, AnyBody Support
> > > > > > > > > > >
> > > > > > > > > > > — In
anyscript@yahoogroups.com, “eugeneyang40”
> > > > > > <eugeneyang40@>
> > > > > > > > > > wrote:
> > > > > > > > > > > >
> > > > > > > > > > > > Hi all,
> > > > > > > > > > > >
> > > > > > > > > > > > I just started using Anybody about 2 weeks
> ago.
> > I
> > > > went
> > > > > > through
> > > > > > > > > > the
> > > > > > > > > > > > whole set of tutorial already. I am
currently
> a
> > 4th
> > > > yr
> > > > > > > > > > undergraduate at
> > > > > > > > > > > > Nanyang Technological University and i’m
doing
> > my
> > > > Final
> > > > > > Yr
> > > > > > > > > > Project on
> > > > > > > > > > > > Single leg landing.
> > > > > > > > > > > >
> > > > > > > > > > > > I decide to use the Gait3D in the Repository
> as
> > a
> > > > start
> > > > > > to study
> > > > > > > > > > the
> > > > > > > > > > > > Knee Joint Force. Base on the Markers
position
> > of
> > > > Kit
> > > > > > Vaughan
> > > > > > > > > > Markers
> > > > > > > > > > > > as use in Gait3D, I did a Motion Capture
with
> > > > Qualisys
> > > > > > and
> > > > > > > > > > subsitute
> > > > > > > > > > > > the P1 to P15 file. I manage to load the
> model.
> > But
> > > > > when
> > > > > > i wish
> > > > > > > > > > to run
> > > > > > > > > > > > kinematic analysis, it showing above error
> > > tolerance
> > > > > for
> > > > > > the
> > > > > > > > > > joint
> > > > > > > > > > > > constraints. May I know how can i solve this
> > > problem.
> > > > > > > > > > > >
> > > > > > > > > > > > Best Regards
> > > > > > > > > > > > Eugene Yang
> > > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > >
> > > >
> _________________________________________________________________
> > > > > > > http://im.live.com/messenger/im/home/?
> > source=hmtextlinkjuly07
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>