Page 2 of 2

Re: Predict walking pattern by using OpenSim-Moco

Posted: Tue Aug 25, 2020 4:42 am
by 04kulaburak
Dear All,

Firstly, I want to thank you all for your replies and help.

Converge of the matlab script took so long in my computer, although I restricted the max iteration number with 200.

Moreover when I modified Ross Miller's osim model for my study and tried this modified model with squad to stand matlab script (Because I could not find the prediction part of Miller's matlab script). After script run, I obtained only environment of OpenSim Moco Visualizer (only blue and brawn screen).

As a result I could not see my model with other scripts, Do you have any idea to see my model with other codes ?

Best Regards,
Burak.

Re: Predict walking pattern by using OpenSim-Moco

Posted: Fri Aug 28, 2020 3:24 am
by rosshm
Hi Burak,

200 iterations is probably not enough for consistent convergence in my experience. If you are running into impractically long CPU times, you might want to start using a 2-D model. It will be much faster than a 3-D model.

The Matlab code I posted is set up for doing tracking simulations of gait, with "sub-task" MocoGoals that are relevant for gait (e.g. periodicity, average speed) but probably not relevant for sit-to-stand. To do predictive simulations of sit-to-stand you would need to remove the state and GRF tracking problems (or replace the data they are tracking with sit-to-stand data) and then specify new MocoGoals that are relevant for sit-to-stand:

https://opensim-org.github.io/opensim-m ... _goal.html

The SquatToStand example that comes with Moco may be a better starting point than my code for what you're trying to do.

Ross