Possibility of running a full lower extremity simulation with only right leg motion-capture data

[Willem Ruseno]

I had a set-up consisting of 2 action cameras (GoPro). So far, the set-up and the application (self-built) from my biomechanics lab have only been attuned to take only one side of the body (right leg). The use of 4 action cameras to take both sides is still under-development, so i'd rather not use it yet for the moment.

Marker set that we are using is the Helen-Hayes marker set. A total of 7 markers on each side. [ASIS - Hip - Femoral wand - Femoral epicondyle - tibial wand - Malleolus - Metatarsal head]

Obviously, with the missing other half of the marker trajectory data, data manipulation or predictive simulation need to be done, which i am currently not familiar yet. Is there anyone who can help, or is experienced with such matter? Advices are greatly appreciated. This is also because the lab currently only has 1 force plate, only enough for one foot, so only one leg's grf can be measured.

Pertaining to the matter above, I've tried running the Inverse Dynamics tool on Gait2354, and tried deleting the Left GRF, but it immediately crashed when i try to save the GRF xml config file (which is containing right GRF only). The same thing happened when I used Rajagopal et. al. full body model and the experiment data plus the setup files. Is this supposed to happen? Does that mean GRF force plate reading from both legs are required?

Also, another concern is that there are no markers attached to the upper extremity (torso, arms). Is it still possible to run the simulation using models such as gait2354 or gait2359? Because as far as I've searched, there are no models that has only lower extremity body. Locking the coordinates or disabling the markers resulted in the torso swaying (which probably would affect the inverse dynamics and so on due to shifting CG)

Is removing the upper extremities body part can still yield a reliable result? (for inv kinematics - inv dynamics - CMC]

your help is greatly appreciated!

addition: if data manipulation can be done, was there an already existing way to "mirror" the data of the right leg as input for left leg data? i understand they too would need to be "tinkered" with since the swing and stance phase occur at different time.

[Aaron Fox]

Hi Willem,

You might want to check out some of the OpenSim Moco examples and features. There are tools within it that allow you to put symmetry constraints within predictive or tracking simulations that might suit your problem. I believe there are also some examples in it that have a locked torso that remains upright.

Aaron

[Willem Ruseno]

Hi Aaron,

Thank you for the heads up! I'll make sure to check them!


Willem