Introduction and questions about gait simmulation

Archive AnyScript Yahoo archive
1

Hi everybody

I’m Christine Schaerer, a graduand in human movement sciences at the
Swiss Federal Institute (ETH) in Zürich. The topic of my diploma
thesis will be the strain distribution in the femur during gait.

Therefore I want to couple the AnyBody-model with an FE-Model of a
femur. That means, I will use AnyBody to calculate the muscle forces
at specific moments during gait cycle and impose them into my FE-
Model of the femur.

The problems are as follows: I will use the kinematic datas from a
previous study. As CT-Scans from the lower limb of this proband are
not available, my FE-model is built on a CT-Scan of a cadaver. Thus,
the dimensions of the cadaver are different from the ones of the
proband, and these are again different from the ones of the AnyBody
model.

The question is now: what is now the smartest way to drive my model?
My idea was to define the exact position of the ankle according to
the kinematic datas, and then drive the angles between the segments.
This way, I won’t have troubles with the different dimensions of the
segments, will I? Is this practicable and reasonable?
In natural gait, there will also be ab/abduction and rotation in the
ankle and the knee, but these two joints are simulated as hinge
joint. How should I handle these further DoF?

I’ve looked at your gait example and read the comments about the
ground reaction force. But I still don’t understand how to impose the
varying direction and amplitude of the force.

As I want to use my own bone model of the femur I have to do quite a
lot of adjustment to the AnyBody model, i.e. scaling the segment,
adapt the muscle attachment etc.
Hence I have a few questions to the body model: How did you determine
the centre of mass (expecially the x- and z-coordinates (femur,
pelvis)) and the location of the knee joint (is it the midpoint
between lateral and medial epicondyles?)? And how are the segment
coordinate systems set (using the ISB standards and then transferred
to the center of mass?)?

Thanks a lot for your help.

Kind regards

Christine

2

Hello Christine and welcome to the group

— In anyscript@yahoogroups.com, “christine_schaerer”
<christine_schaerer@y…> wrote:
>
>
> Hi everybody
>
> I’m Christine Schaerer, a graduand in human movement sciences at
the
> Swiss Federal Institute (ETH) in Zürich. The topic of my diploma
> thesis will be the strain distribution in the femur during gait.
>
> Therefore I want to couple the AnyBody-model with an FE-Model of a
> femur. That means, I will use AnyBody to calculate the muscle
forces
> at specific moments during gait cycle and impose them into my FE-
> Model of the femur.
>
> The problems are as follows: I will use the kinematic datas from a
> previous study. As CT-Scans from the lower limb of this proband are
> not available, my FE-model is built on a CT-Scan of a cadaver.
Thus,
> the dimensions of the cadaver are different from the ones of the
> proband, and these are again different from the ones of the AnyBody
> model.
>
> The question is now: what is now the smartest way to drive my
model?
> My idea was to define the exact position of the ankle according to
> the kinematic datas, and then drive the angles between the
segments.
> This way, I won’t have troubles with the different dimensions of
the
> segments, will I? Is this practicable and reasonable?

This is also the strategy that we would suggest, using the angles
should reduce the problem of having different dimensions.

> In natural gait, there will also be ab/abduction and rotation in
the
> ankle and the knee, but these two joints are simulated as hinge
> joint. How should I handle these further DoF?
>
In the new respository which will be avaiable shortly the ankle joint
is made as a universal joint. I think you will have to disregard the
rotations of these dof until they are present in the model.

> I’ve looked at your gait example and read the comments about the
> ground reaction force. But I still don’t understand how to impose
the
> varying direction and amplitude of the force.

I have made a new version of the gait example an uploaded it to the
file section with the name Gait2.zip.

In the end of the main file i have added a force to the foot which is
defined by a spline. It does not change direction since it only adds
a force in the x direction, but by using a similar procedure for y
the y and z direction the direction could be changed.

>
> As I want to use my own bone model of the femur I have to do quite
a
> lot of adjustment to the AnyBody model, i.e. scaling the segment,
> adapt the muscle attachment etc.
> Hence I have a few questions to the body model: How did you
determine
> the centre of mass (expecially the x- and z-coordinates (femur,
> pelvis)) and the location of the knee joint (is it the midpoint
> between lateral and medial epicondyles?)? And how are the segment
> coordinate systems set (using the ISB standards and then
transferred
> to the center of mass?)?
>

We indeed used the ISB standard for the local coordination systems.
This means that the Y-axis is in the longitudinal direction of the
segment.
In our model the center of mass is the origin of the segment. The
coordinates of the knee and the hip we found with the help of
information from Winter ( Winter, David A.: Biomechanics and motor
control of human movement. Second Edition. John Wiley and Sons, Inc.
ISBN 0-471-50908-6. 1990. ) in the chapter about anthropometrics.
However with this information you only can find the Y-coordinates.
For the other coordinates we used some approximations plus an .stl
file of a femur. Indeed a good approximation for the knee is the
midpoint between the condyles. The hipjoint coordinate was chosen as
the center of the femur head with the help of the stl file.

Best regards
Søren
Anybody support

> Thanks a lot for your help.
>
> Kind regards
>
> Christine