GaitFullBody Model Elbow twists

Main Forum
1

Hello all,

I’m using the GaitFullBody model and removed the lower limbs and the left arm as I’m only interested in movements with the right arm.

I’m using 3D motion capture data to drive my model and I seem to have a problem with driving the Elbow correctly. The RELB marker seems to twist the Elbow incorrectly and not according to the original .c3d file even though I’ve turned off filtering in AnyBody.

So in short:
The RELB marker is showing movement which is not coherent with the original .c3d data. I’ve turned off filtering in AnyBody. The RELB position read in AnyBody (through AnyChart) does not show any difference with the original .c3d data, yet the movement is not the same.

As attachment, I’ve added the .c3d file and model.

Greetings,

Hans Essers

2

Dear Hans,

  1. Could you please let us know which version of AMS and AMMR you are using right now?

  2. Could you please update your profile(organization) ? Then we can identify where you are working for.

Best regards,
Moonki

3

Hello Moonki,

I`m currently using the AnyBody 5.0 with the AMMRV1.3

Greetings,

Hans

4

Hi Hans,

In your zip file I could see only few files which are not enough to run your model.

You should make a zip file which contains all necessary files of your model. And try to upload here again.

Best regards,
Moonki

5

I’ve added all the files in my model folder. I’m not sure if I modified any other .any files except for the arm segment which I’ve added in the previous zip.

6

Update:

It happens in at least 4 subjects so far. I’ve tried to put OptimizeAnthropometricsOnOff On and Off for the Upper- and LowerArm and the Elbow seems to twist less/more during the optimization.

7

Hi Hans,

Now we are trying to figure out which problems are there in your model.

But in your zip file there is no C3D file which we should test.

I think you should upload a new zip file again.

Best regards,
Moonki

8

I’ve added the c3d file in the first zip file in this thread.

9

Hi Hans,

When I tried to load your model, I could see the following error.

ERROR(SCR.PRS9) : C:/W…E/[…]/A…1/M…l/s…t/M…l/Environment.any : ‘McArmNodeP’ : Unresolved object
Model loading skipped

I found the code which is causing this error:
AnyForce3D McArmCWP =
{
F = .CW/2;
AnyRefFrame &MACWP = Main.Studies.KinematicStudyForParameterIdentification.HumanModel.BodyModel.Right.ShoulderArm.Seg.Ulna.McArmNodeP;
…

It seems that you have defined a new node in the repository. And we don’t have that information.

Generally speaking, it is very hard for us to support some models which uses their own ‘modified’ version of repository. We can’t follow up all individual repositories.

I have another suggestion for you.

Actually, you can define some extra elements inside human model not by touching the repository file itself.

You can even add some extra codes in the main file such as:
Main.Studies.KinematicStudyForParameterIdentification.HumanModel.BodyModel.Right.ShoulderArm.Seg.Ulna =
{
AnyRefFrame McArmNodeP = {…};
…
};

Could you please re-configure your model so which can be used with the normal AMMR repository?

If it is not possible, then you should share whole of your modified AMMR with us by through some file-sharing service such as Dropbox.

Best regards,
Moonki

10

Hmm oke, I can’t adjust it like that today. I’ll try asap.

In the first zip file, I’ve also added the Seg.any which I’ve modified for the hand and Environment.any for any forces on the nodes in the modified section.

The Seg.any originates from AnyBody Technology\AnyBody.5.x.x\AMMRV1.3\Body\AAUHuman\Arm

11

Hi,

Now I could see why you have added that ‘Seg.any’ file in the first zip file.

But as I told already, we can’t manage several versions of modified AMMR for forum users. So we can’t modify our own AMMR with your new Seg.any file.

Please attach a single and perfect model which contains all necessary files and can be run with the normal version of AMMR V1.3.

Best regards,
Moonki

12

Hello,

I’ve removed the 2 nodes, they are not needed for this version. It should work with the standard AMMR V1.3. I’ve also added the .c3d file. The only thing that remains is to adjust the link where the .c3d file is read. It is early in the Main file and I’ve added the comment: //fill in .C3D filename total link without .c3d

Thank you for helping out.

Greetings,

Hans

13

Hi Hans,

Sorry for late reply.

Attached please find a modified version of your model.

I tried to change the marker locations.
Also I tried to optimize the segment lengths including the upper and the lower arms.

Please review this model and let’s discuss again if you can see still some problems.

Best regards,
Moonki

14

Hello Moonki,

Thank you for this modified version. I’ve run some more cases and it seems that the elbow problem is not fixed in all subjects. Unfortunately another problem has shown up: the scapula seems to start mid-chest for several subjects. It occurred in the version I’ve sent you and the one you’ve sent me. I’ve included one trial in which this occurs (renamed .c3d to .txt for uploading convenience).

Could you take a look at it and if possible explain where I should search if I want to figure out the problem?

Thanks,

Hans

15

Hi Hans,

Sorry for late reply.

Unfortunately, I could not open your file.

Please start a new thread, upload the full model(including the c3d file) which may cause some problems in a zip file format.
Also please write a corresponding AMS and AMMR versions.

Best regards,
Moonki

(Sorry for not helping you quickly because I will be out of office during two weeks)