Trouble with Mannequin File and Kinematics Analysis

Hello Again!
What a wonderful new forum! Congrats and thank you for creating such a user-friendly, aesthetically-pleasing opportunity to share and ask questions.
I am writing because I have been working on what should be a relatively simple problem, however I have not been able to work my way through it. In the past, I haven’t had a hard time getting a WC model to run, but for some reason for this particular individual, I can’t get the kinematics to look correct, nor can I completely drive the right had side via markers. So what I am saying is:
1.) I can completely drive the left side via markers, however I still have a problem regarding the left wrist – even after adjusting the Mannequin file, it still produces physiologically unrealistic motions during the kinematics analysis, even though the analysis is able to finish;
2.) I am unable to completely drive the right side via markers – I cannot drive the GH external rotation degree of freedom with markers without the kinematics analysis stopping at step 73;
3.) The posture is not driven completely either;
4.) In addition, the wrist on the right side as well, appears to move in a physiologically unrealistic way throughout the kinematics analysis. Adjusting the mannequin file has not helped, and I am at a loss!
I have uploaded the files, and I was hoping that someone could take a look at what I’ve done (all files can go in a folder in the Examples folder of the Repository and the links will then be correct). I’m concerned that since I’m having such a tough time with the kinematics analysis that the inverse dynamics analysis may prove impossible!
And just a note on the files – due to our data collection speed, the PointsMarkers.any file was too large to upload, so I cut it into 8 small files. It is easy to just take the first AAPreInt22.PointsMarkers.any file, open it up, and then proceed to cut and paste the other files (in order) into the AAPreInt22.PointsMarkers.any file. Sorry for the inconvenience in this, but believe it or not I have already cut down the files, and am still running over on size. I guess next time I won’t set my acquisition speed so high.
I look forward to hearing back from you and thank you in advance for your help!
Cheers,
Sarah

Hi Sarah

Welcome to the new forum, and thanks for the kind words.

I have been reviewing your JointAndDrivers.any file to see how the model is driven. I have a few suggestions and questions. I have not tested the model yet but reviewed the driver file, you did not state which repository version the model will run with :wink:

  1. The STN marker seem not to be using the segment coordinate system but the global, this may cause problems if the diretion of the motion changes, since you would no longer be driven the x direction but something else, so i would recommend changing this to use the local coordinate system like the rest of the marker driver are doing.

  2. I am worried about the drivers for the PS dof, I think you may risk the for some angles and positions the two marker drivers on US and RS may point in the exact same direction causing problems, they both point in Z direction.

Here are some answers to your questions:
1: The marker on the hand has a short moment arm and the are no drivers that enforce an exact wrist position, this is gicen indirectly by the elbow position. This means that if you lowerarm length is too long or short it will have a big impact on the wrist motion. The remedy is to tune the segments length to match in the best possible way. A simple check on if the scaling is ok can be done be creating an AnyKinPline between your makers on the elbow and US/RS to read out the distance and to compare that with the one of the time frame of the mocap data.

2: I think you should try to remove one marker driver at the time and replace it by an equivalent joint driver, to be able to see whre the problem relies, it is not nessarily the external rotation, that will fix it.

3: i am not sure i understand this but maybee the change of the STN driver will help on this.

4: Adjusting the mannequin file only helps to solve the initial step, i think it is a problem of segment length and marker location wrt. to the segment that are causing the problems. So it is important to adjust the manneqiun file to the get initial position correct but it also may help you to see which of the segment have the wrong length.

Concerning the file size limitation i will try to increase this.

We are currently implementing methods to solve kinematically overdetermined systems, which will make it much faster to solve this kind of problems, in the future.

Best regards
Søren

Hi Søren,

Thanks for your reply! I am sorry I forgot to mention that I am using AnyBody version 3.0 along with Repository 7.0.

I am going to start over regarding driving the model, and make sure that I am driving each segment using segment coordinates. When you mention this, you mean segment coordinate of, for example, the hand, the lower arm, the upper arm, etc. (you don’t mean coordinates of the markers, right?). I will also take a look at some of your other comments and see if I can’t get this up and running!

Thanks again, Søren!
Sarah

Hi Sarah,

The solvers in version 4 have been completely updated, and I think this will be a major improvement to your WC model. The new ones are much more stable than previously and also give you much smoother activation curves. So if you have a chance I very much recommend changing to version 4 of the system.

Best regards,
John