Hello Sylvain,
Thanks one more time for you reply and your support.
The error that did not appear is shown below and I have also attached the video
of the muscle driven model in the current mail.
Â
Error:
‘MyStudy’ : Muscle recruitment analysis failed, simplex solver reports that
solution does not satisfy all constraints.
I will now freeze the Anybody modelling because I am also developing a full
skeletal model using BodyBuilder/Vicon and as soon as I will finish my
exeperiments too, then I will start developing the full marker-driven AnyBody
Model.
Thanks again for your valuable support.
Best wishes
George
----- Original Message ----
From: AnyScript Support <anyscriptsup@anybodytech.com>
To: anyscript@yahoogroups.com
Sent: Monday, October 6, 2008 3:46:00 PM
Subject: RE: [AnyScript] Drivers for a Simplified Rowing Model
Hi George,
By some mistake the error message does not appear in your email (see below,
only blank lines). Well that makes it a bit harder to point at the problem.
It seems from the driver of the handle SegHHMotion that you did not used a
joint to link it to the ground, but you used instead some kinematic
measures. That if fine. However a joint provide by itself reaction forces in
the directions where the motion is constrained. When you are modelling a
joint by kinematic measures as you did you have to remember to recreate
those intrinsic reaction forces of the joint. This is done by switching On
the appropriate components of the Reaction.Type.
For example I understand the handle has 2 degrees of freedom, it can slide
in the anterior/posterior direction and somehow move up and down, is that
right? Well in all cases the components of the motion that you consider as
free should have the Reaction.Type Off. But the other components of the
movement are supposed to be blocked by some mechanical device (if the handle
is sliding on a rail the lateral movement will blocked mechanically by the
rail), so the corresponding component of the Reaction.Type should be On.
So if the handle has 2 degree of freedom, then 2 components of the
Reaction.Type should be Off and the 4 remaining On.
One last thing about the muscles: I could not download the video for some
unknown reasons, so I could not see if there are enough muscles. But you
have to make sure that there are muscles for every degree of freedom of the
body model (and in both directions of each dof, muscles can only pull).
I hope that help.
Best regards,
Sylvain, AnyBody Support
From: anyscript@yahoogrou ps.com [mailto:anyscript@yahoogrou ps.com] On Behalf
Of GEORGIOS CHRISTODOULAKIS
Sent: 4. oktober 2008 14:20
To: anyscript@yahoogrou ps.com
Subject: Re: [AnyScript] Drivers for a Simplified Rowing Model
Hi Sylvain,
Thanks a lot! The model now works. The problem was the AnyReacForce and I
deleted the corresponding code.
I developed a new model in which I attached muscles to the trunk and the
upper extremity.The model works fine and is muscle driven this time.
There is only one problem now. All the drivers except the Ground/Feet have
Reaction.Type = {0}; and when I am running the model I have the following
message:
That means that I need to add more muscles to the model, is that correct?
So my next step was to identify the component of motion that there is
problem. I found out that when I switch on the Reaction.Type at the
translational x or y or the rotational z (at the handle driver) the model
works. And for the reason that I wanted to find out the effect of an
external force at the handle I decided to switch on only the Reaction.Type
of the rotational z. In order also to check if there is an effect of
switching on the Reaction.Type of the z-rotational to the reliability of the
reaction forces at the joints I switched on the Reaction.Type of
x-rotational and y-rotational at the same time and there was no change at
the results. If the model was 100% muscle driven would I have different
results at the joint reaction forces?
I have uploaded a movie of the muscle driven model and the corresponding
drivers (Northumbria Folder).
Thank you in advance.
I am looking forward to hearing from you.
Best wishes
George
RE: [AnyScript] Drivers for a Simplified Rowing Model
Hi George,
The warning message you dumped clearly says that there are still too much
reaction forces in the model. As you can see there are redundant reactions
and it is dealing with the AnyReacForce you added to the feet. I can’t be
sure of it without seeing the kinematic measure you added the AnyReacForce
to, but it seems to me this force is between the feet and the ground (or the
footrest somehow connected to ground). Is that right? Then it seems to me
you already have a joint there (between feet and ground) with a driver with
reaction type “On”. In that case the mechanical reaction forces of the joint
and the reaction forces of the driver are redundant with the AnyReacForce.
And you should delete one of them (I would suggest the AnyReacForce) .
If that still not solve the problem then please upload the entire
application so we can have a closer look at it.
Best regards,
Sylvain, AnyBody Support
From: anyscript@yahoogrou <mailto:anyscript% 40yahoogroups. com> ps.com
[mailto:anyscript@ yahoogrou <mailto:anyscript% 40yahoogroups. com> ps.com] On
Behalf
Of GEORGIOS CHRISTODOULAKIS
Sent: 1. oktober 2008 19:15
To: anyscript@yahoogrou <mailto:anyscript% 40yahoogroups. com> ps.com
Subject: Re: [AnyScript] Drivers for a Simplified Rowing Model
Hello Sylvain,
I am writing to you again because I have problem with a new rowing model I
developed.
This time I added muscles to the model I put Reaction.Type = {0}; at the
rail driver and the model moves.
The inverse dynamic analysis is completed but the muscles are not loaded.
Could you please explain me how the whole assembly moves if the muscles are
not loaded while the Reaction.Type is 0 which means that the muscles have to
do the work.
In addition when I apply a force at the “pelvis” (Seg3) there is no effect
at the joint reaction forces except the force at the z-direction.
I have uploaded some files at the “Northumbria” folder in order to have a
better look to my problem.
Thank you in advance.
Best wishes
George
----- Original Message ----
From: AnyScript Support <anyscriptsup@
<mailto:anyscriptsu p%40anybodytech. com> anybodytech. com>
To: anyscript@yahoogrou <mailto:anyscript% 40yahoogroups. com> ps.com
Sent: Wednesday, October 1, 2008 9:05:04 AM
Subject: RE: [AnyScript] Drivers for a Simplified Rowing Model
Hi George,
Welcome to the group.
From the video the motion seems to be good. It is actually a good choice to
drive the model with the seat and handle as you don’t have precise data for
the upper and lower extremities motion.
The reason why you can’t see any reaction in the joints is because the
reaction type of the drivers is turned On. This means that there is a kind
of artificial force reaction in the joint that will carry all the forces
applied to the joint. So when you apply a force on the handle or the seat,
the driver of the handle or the seat will carry this force, and the body
does not have to carry it. The good boundary conditions would be to have the
feet fixed to the ground with the reaction type On, and switch Off the
reaction type of the handle and seat driver (because the body is supposed to
provide the force to move those two elements).
However it seems from the video that you are building a full body model from
scratch. I don’t know what is your experience with AnyBody but just the task
of building the body model is a very complicated one if you want an accurate
model. Maybe you already know that we have a public library of body models
and applications available to download. Here is the link:
http://anybody. aau.dk/repositor y/index.htm
My recommendation is to use the existing body model but if you choose to
build your own body model you will need to add muscle to it in order to run
the inverse dynamic analysis and compute forces on the body.
There are also some tutorials on our web site (
http://www.anybodyt ech.com/652. 0.html ) that might help you in your task.
About the last question for the wrist joint, what you are doing is correct.
But there is an easier way to do it, there is a joint called
AnyUniversalJoint that only allows two rotations only (you can choose the
axis).
Best regards,
Sylvain, AnyBody Support
From: anyscript@yahoogrou ps.com [mailto:anyscript@ yahoogrou ps.com] On
Behalf
Of christodoulakis
Sent: 27. september 2008 22:16
To: anyscript@yahoogrou ps.com
Subject: [AnyScript] Drivers for a Simplified Rowing Model
Dear AnyBody Team
I am a new member in your group and I am currently studying for a PhD
in Rowing Biomechanics at Northumbria University at Newcastle (UK).
I am currently developing a simplified rowing model before start
developing a full marker-driven model.
I am writing to you because I need your advice please.
The model that I have developed is driven (except the feet drivers)
with three drivers. These are: the seat driver (prismatic joint), the
combinational spine driver and the handle driver. As a result the
lower extremities follow the motion of the seat driver, the spine
segments are all driven, and the upper extremities follow the motion
of the handle. That means that the joints of ankle, knee, hip,
shoulder, elbow and wrist have no drivers.
My problem now is when I apply an external force at the handle or the
seat there is no effect at the joint reaction forces of the lower
extremities and the upper extermities. Is that because there no
drivers at those joints?
If this is the problem then what kind of drivers should I use for the
upper extremity where the motion is complicated and it is difficult
to use a driver for the spherical joint because of the constraint
violations.
For the lower extremity I can disable the driver of the prismatic
joint and put drivers at the ankle joint and the knee joint.
So for the upper extermity should I use some simple muscles that will
give motion to all the segments or there is any other way of driving
the model without muscles?
Is it possible to drive the model with interpolation drivers at each
segment since I can extract the coordinates from the current
configuration which works well?
One last question please. For the wrist joint that there is motion
about the two axes is that correct to use a spherical joint and
constraint the motion about the remaining axis using a driver with
MeasureOrganizer = {1} if for example the axis to be constraint is Y.
In order to make it easier for you to understand how my model works I
have created a folder (Northumbria) which contains a movie of the
model and the Drivers.any file.
Thanks a lot.
I am looking forward to hearing from you.
Best wishes
George
‘MyStudy’ : Muscle recruitment analysis failed, simplex solver reports that
solution does not satisfy all constraints
[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]