Can 3D coordinates drive the human model?

Hello,
I want to simulate skiing movement in AnyBody, but I only have the 3D coordinates of the movement. It is a dat file which contains the 3D coordinates of different body parts at different times. It is easy to extract the data from this file. My idea is to convert dat file to c3d file or use AnyKinEqInterPolDriver, but I don't know how to achieve. Can 3D coordinates data drive the human model? What should I do?
Your help is greatly appreciated.
Best regards.
Zhanpeng Li

Hi Zhanpeng,

If you have a way to convert the data to C3D it will by far the easiest way in terms on AnyBody modeling, because you can use the existing Mocap models, which are bases on C3D files.

Alternatively you can built the model using the AnyKinMotionDriver object which is also used by the Mocap models, it requires an AnyFunInterpol to be constructed for each marker (This part is done automatically when you read a C3D) If you are new to AnyBody and know how to convert to C3D this will be the path of least resistance.

Best regards
Søren

Dear Søren:
Thanks for your advice. Now I am trying to write a C++ program to convert the dat file to C3D file according to the C3D format.
I have another question. Can the Mocap model use parameter studies?
I want to use parameter studies(or ‘inverse-inverse dynamics’ in some literatures) to optimize the skiing movement.The design variable is the skier’s trunk inclination and knee angle during the flight phase. The objective is to minimize the ACL tension when landing on the ground. In the Mocap model, the CType of C3DMotion is soft. Is it mean that the marker on the segement and the marker from the C3D file don’t have to coincide exactly during the movement? Can I use parameter studies to adjust the movement driven by C3D data? Looking forward to your reply.

Best regards.
Zhanpeng Li

Hi Zhanpeng,

This sounds as a rather challenging model to do.

Please note that the ACL is not in the standard AMMR model it has by default a revolute joint in the knee.

Concerning altering the motion i would recommend you to first extract the motion using the makers then you will have all joint angles + pelvic motion... then you can use inverse-inverse to alter this motion.

Please note that to do this successfully you will need to be able to parameterize the motion somehow with few enough design variables i think this will be the challenging part. Then you will need to setup the objective function.

Best regards
Søren

This topic was automatically closed 125 days after the last reply. New replies are no longer allowed.