Hello! I am new to MOCO so perhaps my problem is obvious.
I am tracking markers only (no GRFs) of some of my own data (trials + model), following the exampleMarkerTracking10DOF.m example. However, there seems to be an issue with tracking longer files. I have some trials which are ~12 seconds long and these fail after going into restorative phases for >1000 iterations. However, if I break them up into smaller chunks using the setTimeBounds function (seems to be around the 5 second limit, anything longer fails) then the tracking solves. I have tried experimenting with actuator values and different marker weightings to no avail. All of my markers are visible and have been visualised in OpenSim using Preview Experimental Data and all looks good. I have tried multiple trials and have the issue for all trials. Is there some sort of limit as to how long the input trial can be? Otherwise I cannot think what might be the issue? Any comments/advice welcome Thank you!
-Ashleigh
MOCO tracking duration limit?
- Ashleigh Louise Wiseman
- Posts: 14
- Joined: Thu Aug 15, 2019 12:53 pm
- Nicholas Bianco
- Posts: 1041
- Joined: Thu Oct 04, 2012 8:09 pm
Re: MOCO tracking duration limit?
Hi Ashleigh,
There's no time limit for problems, but problems longer than a gait cycle or two tend to get very large and take a long time to solve. What mesh interval are you using for this problem? If you are simulating over 10 seconds of data, you might be using a coarse mesh grid, which might be why the problem is having trouble satisfying the constraints.
Either way, I would recommend breaking up the problem into smaller chunks like you're currently doing.
-Nick
There's no time limit for problems, but problems longer than a gait cycle or two tend to get very large and take a long time to solve. What mesh interval are you using for this problem? If you are simulating over 10 seconds of data, you might be using a coarse mesh grid, which might be why the problem is having trouble satisfying the constraints.
Either way, I would recommend breaking up the problem into smaller chunks like you're currently doing.
-Nick