Defining a new marker set

[Aya Alwan]

Hello,

I'm working on a class project, and this is my first time using OpenSim. I would like some guidance on how to use OpenSim to run inverse dynamics on Vicon data. I'm planning to use gait model 2354, but with the locations of our marker set and eventually obtain plots of joint moments for different trials. Can you please direct me on how to set up my model using our marker placement (and preferably our naming convention as well)? I used a static pose trial into the scaling tool but the process was taking so long with no results.

As a side note, I was able to convert my Vicon data files into .trc and .mot so no issues there.

[Daniel Kolb]

Hi Aya,

maybe the Tutorial 3 helps? https://simtk-confluence.stanford.edu:8 ... e+Dynamics
It helped me a lot to get a hang of OpenSim in the beginning and explains pretty good how to use the Scaling Tool, Inverse kinematics Tool as well as the Inverse Dynamics Tool.

I would assume this Tutorial will answer all your questions.

Good luck and have fun,
Daniel

[Aya Alwan]

Hi Daniel, I did go through this tutorial before I started experimenting with my own dataset. I'm assuming that renaming the markers would solve my problem. I'm just not sure how to match my marker locations and naming to the gait model provided. I'm hoping it is a simple fix

Thanks for the help!

Aya

[John Davis]

Usually what I do is go into the model's .osim file (or the marker set .xml file, if you use a separate file) with an XML-capable text editor like Notepad++ or Atom, duplicate some of the pre-existing markers, rename them to my naming convention, and set them at some arbitrary new point in space (doesn't really matter where, just far enough away from the original marker that you can see it clearly). Then you can re-load the model in the GUI, and manually adjust the position of the marker using the axes that show up when you click on it. Then just save it as a new model!

[Aya Alwan]

Thanks John! Trying this now

[Aya Alwan]

Now I have another issue- my model bends 90 degrees forward (attached). I tried to flip the coordinate system but still getting the same results in IK tool. What could be the issue?

[Ayman Habib]

Hi Aya,

Before running IK, I would visualize/inspect the experimental data and see if the trajectory looks correct. It is possible that the data has issues e.g. units, or mislabeling so the solver ends up bending the model to match. Also it is not clear if you have data for the torso, if not then the solver can choose any arbitrary configuration for it.

Hope this helps,
-Ayman