OpenSim crashes during ID-Analysis

[Andreas Vollenweider]

So, I tried everything I could think of.
I changed the modell used to the gait2354 example, used a different set of markers (a smaller one), I did the whole analysis in a different directory on a different harddrive and I only specified times bigger than zero for starting points.

Still somehow i can get no ID to run through or even start properly.

The error log now somehow shows "# EXCEPTION_ACCESS_VIOLATION (0xc0000005) at pc=0x058599b2, pid=7820, tid=6496" instead of the usual one.

I also tried using a former version, but I couldn't use the modell made in the newer version or IK or ID wouldn't even start (just froze without realy freezing).

Is there anything else I could try I haven't thought of yet?

Thank you very much,
Andy

[Ayman Habib]

Andy,

Please report a bug and attach your model and all data and setup files necessary to run it and we'll investigate.

-Ayman

[Andreas Vollenweider]

I have sent a bug report. If any files or informations are still missing, please let me know.
Andy

[Ayman Habib]

Andreas,

Other than the fact that your model has Gravity in -Y but has Z up, I was able to run InverseDynamics using the 2.0.1 distribution. I attached the setup files to the bug report. Please close the bug report upon confirmation.

Thanks,
-Ayman

[Kang Li]

When we change the heads for the grf,
after IK, the arrows representing the grfs disappear. Does anyone have the same problem?

[Chris Zirker]

It turns out that the model was indeed the issue. I've successfully used Sam Hamner's running model to run a handful of inverse dynamics simulations which involve applying four loads simultaneously to the hands and feet, and (some of) the results match up almost exactly to previous dynamics calculations we've made using LifeMod. Well done, OpenSim team!

[Ryan Roemmich]

Hello,

OpenSim is crashing during my analysis as well. All bodies have mass and inertia except for our ground. I am not inputting results of an inverse kinematics analysis (we have created our own motion file for a class project and I'm inputting this instead) and we have no external forces. The program will crash about 10 seconds after I click run. Any input would be greatly appreciated. Thanks.

[Ayman Habib]

Ryan,

Many things can go wrong and cause the Analysis to fail (though it should never crash the application). One potential problem is bad model definition where a coordinate is defined that has no effect on the kinematics which end-up practically confusing Simbody to suspect that the mass matrix is singular, can you check that? Otherwise please report a bug and attach all necessary files and we'll investigate. It would be good to get this resolved within the next couple of days before the next bug-fix release is out.

Best,
-Ayman

[Ryan Roemmich]

To my knowledge, all of the coordinates defined in my model are also used in the motion file. I have submitted a bug report and attached both the motion and the model files. Thanks for the help.

-Ryan

[Ryan Roemmich]

Ayman - I have resolved the error (a silly mistake regarding definition of coordinates for a weld joint). Thank you for your help and please disregard the bug report.