Changing the model in Gait3D example

Hi,

I just tried to change the model being TwoLegs\BodyModel_Mus3E.any to
TwoLegs\BodyModel_NoMuscles.any in the Gait3D application in AREP, and
I got the following error:

ERROR : C:\Grad
school\AnyBody\AnyBodyRepository5\Repository.5\BRep\Aalborg\Leg3D\LegCal1.any(12
)
: MusCalSide : Unresolved object.

Any ideas?

I’m including the file in the files section of this group.

Thanks,
Tilak

PS. I’m interested in looking at the forces in the joints of this
model. Is that possible? Specifically I’m interested at the force at
the L5/S1 joint. Is that present in this model? or must I go to the
legs-and-spine model to be able to see it?

Hi Tilak

I have had a look at the AnyGait.main.any you have uploaded. The
reason for this error message is that the compiler can not find a
object used by the muscle calibration. The reason is that you try to
calibrate the muscles on a model with no muscles. So you should
either shift to a model with muscles or remove the calibration from
the model.

The line

#include “…\BRep\Aalborg\BodyModels\TwoLegs\BodyModel_NoMuscl
es.any”

includes a twoleg model with no muscles on

if you instead use the model with muscles on

//#include “…\BRep\Aalborg\BodyModels\TwoLegs\BodyModel_Mus
3E.any”

it would load.

Alternatively if you want to run the model with no muscles on, you
should outcomment the calibration of the muscles which are defined
using these lines:

//This calibration sequence will calibrate the muscles in the legs
// AnyFolder &HumanFolderRef= Main.Model.HumanModel; This is
the //reference to the folder which contains the human model

#include “…\BRep\Aalborg\BodyModels\FullBodyModel\CalibrationS
equence.any”

You may look in the at the reactions in the joints on the legs. As
an example this would be for the knee joint the following location
in the model tree

Main.Study.Model.HumanModel.Right.Leg.Jnt.Knee.Constraints.Reaction.F

The L4/L5 joint forces can not been seen on this model, since it has
no spine. I think it would be not that difficult to include the
spine model into the gait model, but you would need some more marker
data to drive the motion of the upper body.

Here is a rough sketch on how to do this:

1 use the Bodymodels in the SpineTwoLegs directory instead of the
TwoLegs, this means changing the path in the main file to
#include “…\BRep\Aalborg\BodyModels\SpineTwoLegs\BodyModel_Mu
s3E.any”

  1. You then need to add drivers for the spine motion. Since the
    model automatically make use of a spine rhythm there is three
    rotational dof left free which should somehow be driven. These free
    dof in the spine are the rotations between pelvis and thorax.
    Initial as the first estimation you may copy the driver named
    PostureDriver from the JointsAndDrivers.any file in the
    StandingModel. This will make the spine stiff, but once this is
    running you may release one dof at the time and add more realistic
    drivers, using marker data.

  2. Final step is to move the AnyReacForce attached to the pelvis to
    be attached onto thorax, this is done in the JointsAndDrivers.any
    file.

  3. If you have not done it already, please use the newest repository
    as a basis for the model. It was uploaded last week.

If you have the data for the motion of the upperbody i think it
would be a nice model.

Please do not hesitate to ask again if you have further questions.

Best regards
AnyBody Support

— In anyscript@yahoogroups.com, “duttatil” <tilak.dutta@u…>
wrote:
> Hi,
>
> I just tried to change the model being TwoLegs\BodyModel_Mus3E.any
to
> TwoLegs\BodyModel_NoMuscles.any in the Gait3D application in AREP,
and
> I got the following error:
>
> ERROR : C:\Grad
> school\AnyBody\AnyBodyRepository5\Repository.5
\BRep\Aalborg\Leg3D\LegCal1.any(12)
> : MusCalSide : Unresolved object.
>
> Any ideas?
>
> I’m including the file in the files section of this group.
>
> Thanks,
> Tilak
>
> PS. I’m interested in looking at the forces in the joints of this
> model. Is that possible? Specifically I’m interested at the force
at
> the L5/S1 joint. Is that present in this model? or must I go to the
> legs-and-spine model to be able to see it?

> The L4/L5 joint forces can not been seen on this model, since it has
> no spine.
>

I’m actually interested in forces at the L5/S1 joint. It seems that this may
still be accessible even without a spine attached as this would be the joint
where the pelvis attaches to the spine. (I’m no expert at anatomy and very well
could be wrong).

(There must be some reaction force applied (perhaps by the environment)to the
upper-most part of the pelvis to keep the body at equilibrium?)

Tilak

> — In anyscript@yahoogroups.com, “duttatil” <tilak.dutta@u…>
> wrote:
> > Hi,
> >
> > I just tried to change the model being TwoLegs\BodyModel_Mus3E.any
> to
> > TwoLegs\BodyModel_NoMuscles.any in the Gait3D application in AREP,
> and
> > I got the following error:
> >
> > ERROR : C:\Grad
> > school\AnyBody\AnyBodyRepository5\Repository.5
> \BRep\Aalborg\Leg3D\LegCal1.any(12)
> > : MusCalSide : Unresolved object.
> >
> > Any ideas?
> >
> > I’m including the file in the files section of this group.
> >
> > Thanks,
> > Tilak
> >
> > PS. I’m interested in looking at the forces in the joints of this
> > model. Is that possible? Specifically I’m interested at the force
> at
> > the L5/S1 joint. Is that present in this model? or must I go to the
> > legs-and-spine model to be able to see it?

Hi Tilak

I am sorry, for some reason i thought it was the L4/L5 joint you
meant.

You are right that to some extend you may access S1/l5 reactions in
the current model.

If you in the file JointsAndDriver.any make a small change like
below where i have inserted the L5PelvisJnt node in the
AnyReacForce, you will get to some extend the reaction in this
joint.

AnyReacForce PelvisGroundReaction={

 AnyKinLinear Lin={
   AnyFixedRefFrame &ref1 =..GlobalRef;
   AnySeg &ref2 =

Main.Model.HumanModel.Trunk.SegmentsLumbar.PelvisSeg.L5PelvisJntNode;
};

 AnyKinRotational Rot={
   AnyFixedRefFrame &ref1 =..GlobalRef;
   AnySeg &ref2 =

Main.Model.HumanModel.Trunk.SegmentsLumbar.PelvisSeg.L5PelvisJntNode;
Type=RotVector;
};
};

The reaction which can be seen in this AnyReacForce is not exactly
the joint force in S1/L5. The reason is that in this model the spine
muscles are missing and this AnyReacforce simply try to balances all
the forces which in reality should have acted on the pelvis. These
forces include both muscles forces and joint forces. So the reaction
you will see in the model will be a mix of those forces.

If you included the spine model with all its muscles spanning this
joint it would be another reaction you would get.

Best regards
AnyBody Support

— In anyscript@yahoogroups.com, tilak.dutta@u… wrote:
>
>
> > The L4/L5 joint forces can not been seen on this model, since it
has
> > no spine.
> >
>
> I’m actually interested in forces at the L5/S1 joint. It seems
that this may
> still be accessible even without a spine attached as this would be
the joint
> where the pelvis attaches to the spine. (I’m no expert at anatomy
and very well
> could be wrong).
>
> (There must be some reaction force applied (perhaps by the
environment)to the
> upper-most part of the pelvis to keep the body at equilibrium?)
>
> Tilak
>
> > — In anyscript@yahoogroups.com, “duttatil” <tilak.dutta@u…>
> > wrote:
> > > Hi,
> > >
> > > I just tried to change the model being
TwoLegs\BodyModel_Mus3E.any
> > to
> > > TwoLegs\BodyModel_NoMuscles.any in the Gait3D application in
AREP,
> > and
> > > I got the following error:
> > >
> > > ERROR : C:\Grad
> > > school\AnyBody\AnyBodyRepository5\Repository.5
> > \BRep\Aalborg\Leg3D\LegCal1.any(12)
> > > : MusCalSide : Unresolved object.
> > >
> > > Any ideas?
> > >
> > > I’m including the file in the files section of this group.
> > >
> > > Thanks,
> > > Tilak
> > >
> > > PS. I’m interested in looking at the forces in the joints of
this
> > > model. Is that possible? Specifically I’m interested at the
force
> > at
> > > the L5/S1 joint. Is that present in this model? or must I go
to the
> > > legs-and-spine model to be able to see it?
>

> Alternatively if you want to run the model with no muscles on, you
> should outcomment the calibration of the muscles which are defined
> using these lines:
>
> //This calibration sequence will calibrate the muscles in the legs
> // AnyFolder &HumanFolderRef= Main.Model.HumanModel; This is
> the //reference to the folder which contains the human model
>
> #include “…\BRep\Aalborg\BodyModels\FullBodyModel\CalibrationS
> equence.any”

So, I tried this and it doesn’t appear to work. I got an error.

ERROR : C:\Grad
school\AnyBody\AnyBodyRepository5\Repository.5\ARep\Toronto\Gait3D\EMGComparison
.any(10)
: Activity : Folder identifier used as expresion argument.

Deleting last loaded model…
Model loading skipped

Any help would be greatly appreciated.

The file is posted in the yahoo group “files” section.

Thanks,
Tilak

Hi Tilak

The object that it can not find is a folder which is used for
comparison of muscle activities in the model and measured EMG. Since
the model now has no muscles the compiler can not find the muscles
needed for the comparison.

Please outcomment the line EMGComparison.any in the Gait.main.any
file, it is in the study folder, then hopefully it will work.

Best regards
AnyBody Support

— In anyscript@yahoogroups.com, tilak.dutta@u… wrote:
>
> > Alternatively if you want to run the model with no muscles on,
you
> > should outcomment the calibration of the muscles which are
defined
> > using these lines:
> >
> > //This calibration sequence will calibrate the muscles in the
legs
> > // AnyFolder &HumanFolderRef= Main.Model.HumanModel; This is
> > the //reference to the folder which contains the human model
> >
> >
#include “…\BRep\Aalborg\BodyModels\FullBodyModel\CalibrationS
> > equence.any”
>
> So, I tried this and it doesn’t appear to work. I got an error.
>
>
> ERROR : C:\Grad
> school\AnyBody\AnyBodyRepository5\Repository.5
\ARep\Toronto\Gait3D\EMGComparison.any(10)
> : Activity : Folder identifier used as expresion argument.
>
> Deleting last loaded model…
> Model loading skipped
>
> Any help would be greatly appreciated.
>
> The file is posted in the yahoo group “files” section.
>
> Thanks,
> Tilak
>

> 2. You then need to add drivers for the spine motion. Since the
> model automatically make use of a spine rhythm there is three
> rotational dof left free which should somehow be driven. These free
> dof in the spine are the rotations between pelvis and thorax.
> Initial as the first estimation you may copy the driver named
> PostureDriver from the JointsAndDrivers.any file in the
> StandingModel. This will make the spine stiff, but once this is
> running you may release one dof at the time and add more realistic
> drivers, using marker data.
>
> 3. Final step is to move the AnyReacForce attached to the pelvis to
> be attached onto thorax, this is done in the JointsAndDrivers.any
> file.
>
> 4. If you have not done it already, please use the newest repository
> as a basis for the model. It was uploaded last week.
>
> If you have the data for the motion of the upperbody i think it
> would be a nice model.

Yay!, It worked. I now have the legs-with-spine-no-muscles model working. (I
have not yet added the additional driver so the spine droops over)

2 questions at this point:

  1. I’m trying to get my head around the “degrees of freedom” constraint. I
    think I understand the concept of dof, but in the case where I am driving a
    model with a “real” set of data, why should “overspecifying” drivers be a
    problem? For example, even if I happened to specify 2 drivers that drive a
    certain segment in the x direction; it seems to me this should be fine as long
    as the two drivers don’t conflict. Since our data comes from actual motion
    capture data - I seem to think that all the motion tracker markers could be
    used to drive 3 dofs.

  2. How can I add a driver at C7 on this model? I understand I need to simply
    add a driver, but do I need to create a marker at this postion or is there one
    already there?

Many thanks,
Tilak

Hi Tilak

I am glad it worked out.

You can not use all the data from the motion markers, this will add
to many dof. to your drivers.

When you have the data from the motion capture system it is certain
that due to skin motion, wobbly masses etc. that the marker drivers
will be in conflict with each other, if you use all the dof. The
relative motion between the markers on the same segment will be much
higher than the tolerances set by the kinematic solver.

So the next step is to remove the rotational driver you added on the
spine, and replace this with marker drivers, like on the legs. When
you do this, it is a good idea to remove the driver for one
rotational dof. at the time, and add one marker driver. This will
make it easier to see the effect of your changes. In the posture
driver you may add a MeasureOrganizer line to control which dof of
the rotational measure you are driving.

There is no point yet for C7 so your are right that this has to be
added to the model. Please add the node in the following file
MarkerPosOnBody.any, simply add a AnyRefNode there. When you make
the node you may add the following line inside the AnyRefNode

AnyDrawRefFrame drw={RGB={0,0,1};};

This will draw a blue coordinate system in the node and in this way
make it easier for you to see its location, and get its position
right.

Best regards
AnyBody Support

— In anyscript@yahoogroups.com, tilak.dutta@u… wrote:
>
>
> > 2. You then need to add drivers for the spine motion. Since the
> > model automatically make use of a spine rhythm there is three
> > rotational dof left free which should somehow be driven. These
free
> > dof in the spine are the rotations between pelvis and thorax.
> > Initial as the first estimation you may copy the driver named
> > PostureDriver from the JointsAndDrivers.any file in the
> > StandingModel. This will make the spine stiff, but once this is
> > running you may release one dof at the time and add more
realistic
> > drivers, using marker data.
> >
> > 3. Final step is to move the AnyReacForce attached to the pelvis
to
> > be attached onto thorax, this is done in the
JointsAndDrivers.any
> > file.
> >
> > 4. If you have not done it already, please use the newest
repository
> > as a basis for the model. It was uploaded last week.
> >
> > If you have the data for the motion of the upperbody i think it
> > would be a nice model.
>
> Yay!, It worked. I now have the legs-with-spine-no-muscles model
working. (I
> have not yet added the additional driver so the spine droops over)
>
> 2 questions at this point:
>
> 1) I’m trying to get my head around the “degrees of freedom”
constraint. I
> think I understand the concept of dof, but in the case where I am
driving a
> model with a “real” set of data, why should “overspecifying”
drivers be a
> problem? For example, even if I happened to specify 2 drivers that
drive a
> certain segment in the x direction; it seems to me this should be
fine as long
> as the two drivers don’t conflict. Since our data comes from
actual motion
> capture data - I seem to think that all the motion tracker markers
could be
> used to drive 3 dofs.
>
> 2) How can I add a driver at C7 on this model? I understand I need
to simply
> add a driver, but do I need to create a marker at this postion or
is there one
> already there?
>
> Many thanks,
> Tilak
>

> When you have the data from the motion capture system it is certain
> that due to skin motion, wobbly masses etc. that the marker drivers
> will be in conflict with each other, if you use all the dof. The
> relative motion between the markers on the same segment will be much
> higher than the tolerances set by the kinematic solver.

Ah…ok. I get it. Thanks.

> So the next step is to remove the rotational driver you added on the
> spine, and replace this with marker drivers, like on the legs.

I need to back up a step. I haven’t yet put a driver on the spine. I assume
what you mean is to do the following:

> 2. You then need to add drivers for the spine motion. Since the
> model automatically make use of a spine rhythm there is three
> rotational dof left free which should somehow be driven. These free
> dof in the spine are the rotations between pelvis and thorax.
> Initial as the first estimation you may copy the driver named
> PostureDriver from the JointsAndDrivers.any file in the
> StandingModel. This will make the spine stiff, but once this is
> running you may release one dof at the time and add more realistic
> drivers, using marker data.

OK. I copied the following from the standing model JointsAndDrivers.any file
into the Gait3D JointsAndDrivers.any file:

AnyKinEqSimpleDriver PostureDriver ={
AnyKinMeasureOrg &Ref2
= …HumanModel.Interface.Trunk.PelvisThoraxLateralBending;
AnyKinMeasureOrg &Ref3
= …HumanModel.Interface.Trunk.PelvisThoraxRotation;

 AnyKinMeasureOrg &Ref1 = ...HumanModel.Interface.Trunk.PelvisThoraxFlexion;

DriverPos = pi/180*{
.JntPos.PelvisThoraxLateralBending,
.JntPos.PelvisThoraxRotation,
.JntPos.PelvisThoraxFlexionExtension
};

 // This is static posture, so all velocities are zero.
 DriverVel = pi/180*{
   0.0,
   0.0,
   0.0
 };

 Reaction.Type = {
   0.0,
   0.0,
   0.0
 };

};

and I got the following error:

ERROR : C:\Grad school\AnyBody\AnyBodyRepository5\Repository.5
\ARep\Toronto\Gait3D\JointsAndDrivers.any(81) : JntPos : Unresolved
object.

Deleting last loaded model…
Model loading skipped

I’m copying the files to the yahoo group “files” section

Thanks for your help,
Tilak

PS. Once we relsolve this, my next question will be if C7 coordinates can be
used to drive the rotational degrees of freedom. (C7 x,y,z won’t give any info
about twisting, of the thorax) Will I need coords for say, both shoulders
instead?

> So the next step is to remove the rotational driver you added on the
> spine, and replace this with marker drivers, like on the legs. When
> you do this, it is a good idea to remove the driver for one
> rotational dof. at the time, and add one marker driver. This will
> make it easier to see the effect of your changes. In the posture
> driver you may add a MeasureOrganizer line to control which dof of
> the rotational measure you are driving.
>
> There is no point yet for C7 so your are right that this has to be
> added to the model.

> There is no point yet for C7 so your are right that this has to be
> added to the model. Please add the node in the following file
> MarkerPosOnBody.any, simply add a AnyRefNode there.

Hi Talak

Yes you were right to copy the posture driver from the standingmodel.
The JntPos, that it can not find is a folder reference that it used
in standing model which gets all its joints angles from a mannequin
file. In this model you may just type in the Driverpos numbers
directly for these drivers.

Such as

AnyKinEqSimpleDriver PostureDriver ={
  AnyKinMeasureOrg &Ref2

= …HumanModel.Interface.Trunk.PelvisThoraxLateralBending;
AnyKinMeasureOrg &Ref3
= …HumanModel.Interface.Trunk.PelvisThoraxRotation;

  AnyKinMeasureOrg

&Ref1 …HumanModel.Interface.Trunk.PelvisThoraxFlexion;
DriverPos = pi/180*{
0,
0,
0 };

  // This is static posture, so all velocities are zero.
  DriverVel = pi/180*{
    0.0,
    0.0,
    0.0
  };

  Reaction.Type = {
    0.0,
    0.0,
    0.0
  };
};

This will set the spine in its natural position.

You are right that a C7 marker driver may be used for driving two of
the rotational dof of the spine and that it can not be used for
driving the twist. It is a good idea to use markers on the shoulder
(these will also have to added to the model). I would suggest adding
the following drivers.

On both shoulders drive the markers in the x direction only.
On C7 driver the sideways direction. This setup will add three
constraints to the poblem which is what you need to be able to
remove the posture driver.

Best regards
AnyBody Support

> > So the next step is to remove the rotational driver you added on
the
> > spine, and replace this with marker drivers, like on the legs.
>
> I need to back up a step. I haven’t yet put a driver on the spine.
I assume
> what you mean is to do the following:
>
> > 2. You then need to add drivers for the spine motion. Since the
> > model automatically make use of a spine rhythm there is three
> > rotational dof left free which should somehow be driven. These
free
> > dof in the spine are the rotations between pelvis and thorax.
> > Initial as the first estimation you may copy the driver named
> > PostureDriver from the JointsAndDrivers.any file in the
> > StandingModel. This will make the spine stiff, but once this is
> > running you may release one dof at the time and add more
realistic
> > drivers, using marker data.
>
> OK. I copied the following from the standing model
JointsAndDrivers.any file
> into the Gait3D JointsAndDrivers.any file:
>
> AnyKinEqSimpleDriver PostureDriver ={
> AnyKinMeasureOrg &Ref2
> = …HumanModel.Interface.Trunk.PelvisThoraxLateralBending;
> AnyKinMeasureOrg &Ref3
> = …HumanModel.Interface.Trunk.PelvisThoraxRotation;
>
> AnyKinMeasureOrg &Ref1
= …HumanModel.Interface.Trunk.PelvisThoraxFlexion;
>
> DriverPos = pi/180*{
> .JntPos.PelvisThoraxLateralBending,
> .JntPos.PelvisThoraxRotation,
> .JntPos.PelvisThoraxFlexionExtension
> };
>
> // This is static posture, so all velocities are zero.
> DriverVel = pi/180*{
> 0.0,
> 0.0,
> 0.0
> };
>
> Reaction.Type = {
> 0.0,
> 0.0,
> 0.0
> };
> };
>
>
> and I got the following error:
>
> ERROR : C:\Grad school\AnyBody\AnyBodyRepository5\Repository.5
> \ARep\Toronto\Gait3D\JointsAndDrivers.any(81) : JntPos :
Unresolved
> object.
>
> Deleting last loaded model…
> Model loading skipped
>
> I’m copying the files to the yahoo group “files” section
>
> Thanks for your help,
> Tilak
>
> PS. Once we relsolve this, my next question will be if C7
coordinates can be
> used to drive the rotational degrees of freedom. (C7 x,y,z won’t
give any info
> about twisting, of the thorax) Will I need coords for say, both
shoulders
> instead?
>
> > So the next step is to remove the rotational driver you added on
the
> > spine, and replace this with marker drivers, like on the legs.
When
> > you do this, it is a good idea to remove the driver for one
> > rotational dof. at the time, and add one marker driver. This
will
> > make it easier to see the effect of your changes. In the posture
> > driver you may add a MeasureOrganizer line to control which dof
of
> > the rotational measure you are driving.
> >
> > There is no point yet for C7 so your are right that this has to
be
> > added to the model.
>
> > There is no point yet for C7 so your are right that this has to
be
> > added to the model. Please add the node in the following file
> > MarkerPosOnBody.any, simply add a AnyRefNode there.
>

>In this model you may just type in the Driverpos numbers
> directly for these drivers.

I did as you suggested and find that while it does compile, I get the following
warnings. Should I be worried about any of this:

Next ERROR is warning now, but will be real ERROR SOON!
ERROR : C:\Grad
school\AnyBody\AnyBodyRepository5\Repository.5\BRep\Aalborg\BodyModels\SpineNoMu
sclesTwoLegs\BodyModel_NoMuscles.any(36)
: PsoasMajorRightReference : Reference of class : AnyRefNode can not refer
to object class : AnySeg. Reference must be same class or parent for referentor.
Next ERROR is warning now, but will be real ERROR SOON!
ERROR : C:\Grad
school\AnyBody\AnyBodyRepository5\Repository.5\BRep\Aalborg\BodyModels\SpineNoMu
sclesTwoLegs\BodyModel_NoMuscles.any(37)
: PsoasMajorLeftReference : Reference of class : AnyRefNode can not refer
to object class : AnySeg. Reference must be same class or parent for referentor.
Evaluating constants…
Configuring model…
Model Warning: Study ‘Main.Study’ contains too few kinematic constraints to be
kinematically determinate.
Loaded successfully.
Elapsed Time : 4.734000

> On both shoulders drive the markers in the x direction only.
> On C7 driver the sideways direction. This setup will add three
> constraints to the poblem which is what you need to be able to
> remove the posture driver.

I’m not positive which direction will end up being “sideways” for our model. (It
depends on how the calibration cube is set up when we collect our motion capture
data. Does this matter? I’m assuming you mean the shoulders drive the forward
direction (in the 3D gait model this is x), and C7 drives y.

If it turns out that our nurse in their activity ends up facing in the y
direction, will I then have to change things so C7 drives the x direction and
shoulders drive y.

Also, would you know of any public access database that might list the pelvic
width, thigh length, shank length, foot length for males and females as a
function of height? I’m trying to find an easy way to estimate these values for
the AnyNurse (same as the AnyMan) file. I think I have a source for everything
but pelvic width - which I was told might be the distance between hip joints.

Thanks so much for your continued assistance,
Tilak

Hi Tilak

To avoid the warning please make the folwoing changes in
the “BRep/aalborg/Bodymodels/SpineNoMusclesTwolegs/Bodymodel_nomuscle
s.any” file

Find these lines
AnyRefNode &PsoasMajorRightReference=.Right.Leg.Seg.Thigh;
AnyRefNode &PsoasMajorLeftReference=.Left.Leg.Seg.Thigh;

and change them to these
AnySeg &PsoasMajorRightReference=.Right.Leg.Seg.Thigh;
AnySeg &PsoasMajorLeftReference=.Left.Leg.Seg.Thigh;

This warning will have no effect on the model yet, but in future
versions it will not be able to load. The error was that the
reference Psoasmajorrightreference was pointing on the thigh segment
and it was constructed as reference node.

By the way i can see that you are using the SpineNoMusclesTwoLegs
model. This model do not have muscles in the spine, it uses joint
reactions in the joints instead of muscles. This will give you
another reaction force in the spine joints than if you use the spine
with the muscles on in the SpineTwoLegs directory, i think you
should use the latter.

The next warning that it displays says that the model has too few
kinematic constraints maybee you have not added the posture driver
yet?

It does not realy matter how the directions ends up, you just need
to change the model drivers accordingly. When i wrote that C7 should
be driven sideways i meant the y direction you were right.

You are right that the pelvis width is the distance between the hip
joints. I do not know of any databases giving the pelvis width as a
function of height, i think the best thing will be to estimate it as
a function of the distance between the trochanter.

Best regards
AnyBody Support

— In anyscript@yahoogroups.com, tilak.dutta@u… wrote:
>
> >In this model you may just type in the Driverpos numbers
> > directly for these drivers.
>
> I did as you suggested and find that while it does compile, I get
the following
> warnings. Should I be worried about any of this:
>
> Next ERROR is warning now, but will be real ERROR SOON!
> ERROR : C:\Grad
> school\AnyBody\AnyBodyRepository5\Repository.5
\BRep\Aalborg\BodyModels\SpineNoMusclesTwoLegs\BodyModel_NoMuscles.an
y(36)
> : PsoasMajorRightReference : Reference of class : AnyRefNode
can not refer
> to object class : AnySeg. Reference must be same class or parent
for referentor.
> Next ERROR is warning now, but will be real ERROR SOON!
> ERROR : C:\Grad
> school\AnyBody\AnyBodyRepository5\Repository.5
\BRep\Aalborg\BodyModels\SpineNoMusclesTwoLegs\BodyModel_NoMuscles.an
y(37)
> : PsoasMajorLeftReference : Reference of class : AnyRefNode
can not refer
> to object class : AnySeg. Reference must be same class or parent
for referentor.
> Evaluating constants…
> Configuring model…
> Model Warning: Study ‘Main.Study’ contains too few kinematic
constraints to be
> kinematically determinate.
> Loaded successfully.
> Elapsed Time : 4.734000
>
> > On both shoulders drive the markers in the x direction only.
> > On C7 driver the sideways direction. This setup will add three
> > constraints to the poblem which is what you need to be able to
> > remove the posture driver.
>
> I’m not positive which direction will end up being “sideways” for
our model. (It
> depends on how the calibration cube is set up when we collect our
motion capture
> data. Does this matter? I’m assuming you mean the shoulders drive
the forward
> direction (in the 3D gait model this is x), and C7 drives y.
>
> If it turns out that our nurse in their activity ends up facing in
the y
> direction, will I then have to change things so C7 drives the x
direction and
> shoulders drive y.
>
> Also, would you know of any public access database that might list
the pelvic
> width, thigh length, shank length, foot length for males and
females as a
> function of height? I’m trying to find an easy way to estimate
these values for
> the AnyNurse (same as the AnyMan) file. I think I have a source
for everything
> but pelvic width - which I was told might be the distance between
hip joints.
>
> Thanks so much for your continued assistance,
> Tilak
>