Static Optimization Issue

[Alex Landauer]

I am experiencing an issue when running the Static Optimization tool, it gets stuck right after printing "Executing the analyses from -2e-007 to 0.99999..." to the message window (or the command line - I've tried both), and the analysis progress bar stays at 0%. To be clear, I receive no messages after "Executing the analyses...", the program does not crash or close, and it consistently uses about 50% CPU, whether running in the GUI or the command line. Its running on an AMD Athlon II X2 245 processor with Windows 7 x64 and 4.00GB RAM. To set up the tool, I ran IK on my marker data and used that motion for the input file, there is no external load, a 6Hz low-pass filter is applied and the "Use muscle force-length-velocity is check, all other values are the defaults. Does anyone have ideas as to what might be causing the hang-up?

Thanks,
Alex Landauer

[Ayman Habib]

Hi Alex,

It's possible that the problem is caused by having way too many degrees of freedom so the distribution problem is harder. Are there coordinates that you can lock to help the optimizer? I don't think we should hang any way so please file a bug report and attach all necessary files to reproduce the problem and we'll investigate.

Good luck,
-Ayman

[Alex Landauer]

Thanks Ayman, by locking coordinates and thereby reducing the number of degrees of freedom, as you suggested, the Static Optimization completed. Now I've moved on to CMC and have persistently been getting an error:

[quote]
Setting optimizer convergence criterion to 1e-006.
Setting optimizer maximum iterations to 100.

Setting cmc controller to not use verbose printing.

This application has requested the Runtime to terminate it in an unusual way.
Please contact the application's support team for more information.



================================================================
================================================================
Computing initial values for muscles states (activation, length)
Start time = Fri Jul 01 10:19:32 2011

================================================================
[/quote]

And the program crashes. Any ideas?

Thanks,
Alex Landauer

[Ayman Habib]

Hi Alex,

It'd be hard to debug this problem remotely, but here are some questions/suggestions:
1. What version of OpenSim are you using and what model?
2. Did you run RRA? The settings should be pretty similar to CMC (in terms of tracking kinematics and External Loads).
3. Did you check the compute-equilibrium box? What integrator tolerance etc.

-Ayman

[Alex Landauer]

Ayman,
Sorry for the long pause. I am using OpenSim 2.2.1 with the Delft Shoulder and Elbow Model by Blana et al. I did not run RRA, and at this point I am have not placed any external loads on the model. I have left everything but "Reduce Residuals" as the default values (on all panes), however I unchecked both "Adjust model" and "Adjust kinematics". My motion file seems to have worked amicably with ID and Static Opt so perhaps the Tasks file I made for CMC is at fault? It is based heavily on the example in the user guide.

On an unrelated note, what is the difference between "DAS3_Kinematics_q.sto" and "DAS3_Kinematics_u.sto", that I get when running the Analysis tool with the kinematics option. I see they have different data, but why?

Thank you,
Alex Landauer

[Ayman Habib]

Hi Alex,

I'd suggest you try 2.3.1 (brya) as it has the RRA ad CMC tools separated and so they each have fewer knobs to tweak.

When you run a Kinematics Analysis you get coordinates (_q) and speeds (_u). Some other analyses produce _states files that contain both coordinates and speeds (along with muscle states).

Hope this explains and good luck,
-Ayman

[Alex Landauer]

Ayman,
Thank you for your prompt reply, your explanation of the “_q” and “_u” data files is completely satisfactory: it explains the results I was seeing in those files well.

As for the CMC issue, I have tried the 2.3.1 version, and it seems like a better setup, but I am still receiving basically the same error when it runs. I get the same output to the message window, but now the tool crashes with the “Tool execution failed. Check massage window for details.” followed immediately by “Tool execution failed or was canceled by user. Output files were not written.” error. I am using a motion file that I adapted from that found in the Arm26 example and am using only three markers (I intend to add more in future). I have altered the file so that the markers do not move; could this be causing the issue? My tasks file is patterned on that found in the User Guide; it tracks a single joint object with weight of 1.0, and a single associated coordinate. Do you (or anyone) have thoughts about what might be causing this?

Thanks,
Alex Landauer