Page 1 of 2
ID problems in 3.0
Posted: Sun Oct 28, 2012 2:30 am
by fausto_p
Hi,
after running scaling and IK I try to run ID but the simulation does not start and I get the following error.
I believe all my inputs are correct, having checked them with those available on line.
Any idea on what's wrong?
Thanks
Re: ID problems in 3.0
Posted: Sun Oct 28, 2012 9:29 pm
by saxbyd
What's the error mate?
Re: ID problems in 3.0
Posted: Sun Oct 28, 2012 9:32 pm
by fausto_p
Sorry my attachment did not go through.
I am posting down here:
Constructing tool from setup file idsetok.xml.
-----------------------------------------------------------------------
Starting Inverse Dynamics
-----------------------------------------------------------------------
-----------------------------------------------------------------------
Loaded model Malcolm_Mintz from file Model with new settings.osim
MODEL: MM
coordinates: 29
forces: 92
actuators: 92
muscles: 92
analyses: 0
probes: 0
bodies: 13
joints: 12
constraints: 0
markers: 32
controllers: 0
contact geometries: 0
misc modelcomponents: 0
Running tool MM.
Storage: file=IKoutput.mot (nr=2285 nc=30)
Low-pass filtering coordinates data with a cutoff frequency of 6...
Storage: file=C:\Documents and Settings\Walk 0,75 m per sec_grf.mot (nr=45740 nc=19)
Exception in ID: SimTK Exception thrown at SmallMatrixMixed.h:878:
Error detected by Simbody method lapackInverse(Mat<>): Matrix is singular so can't be inverted (Lapack getrf info=3).
(Required condition 'info==0' was not met.)
Re: ID problems in 3.0
Posted: Sun Oct 28, 2012 9:41 pm
by aymanh
Hi Fausto,
Are you sure you have mass properties specified for all the segments of the model you're analyzing? These would be needed for InverseDynamics but not scale or IK?
Please let us know so we can help you troubleshoot further.
Best regards,
-Ayman
Re: ID problems in 3.0
Posted: Sun Oct 28, 2012 9:59 pm
by fausto_p
Hi Ayman,
I believe I have mass properties for all my segments (except ground but it should not matter if I am right).
With properties I mean mass, mass center and inertia.
Thanks for any support you could give me.
Fausto
Re: ID problems in 3.0
Posted: Mon Oct 29, 2012 10:01 am
by aymanh
Hi Fausto,
Please file a bug report here (
https://simtk.org/tracker/?atid=322&gro ... unc=browse) and attach the model and all necessary files to reproduce the problem and we'll investigate. Just for reference, are you applying ExternalLoads to your model? IF so, does InverseDynamics work without them? Also did you run these files in the past (in version 2.4 for example)?
Best regards,
-Ayman
Re: ID problems in 3.0
Posted: Wed Mar 20, 2013 12:06 pm
by hsynclk
Hi:
We're getting the same error message. Is this problem solved? Thanks.
Huseyin
Re: ID problems in 3.0
Posted: Thu Mar 21, 2013 7:54 am
by eskilan
I had the same "matrix is singular" problem but running forward dynamics. The problem was that I had muscles in my model but hadn't specified a control for them. Removing the unused muscles solved the problem.
Re: ID problems in 3.0
Posted: Tue May 28, 2013 11:24 am
by jhb777
I'm also getting the same problem with inverse dynamics? What was the solution to this problem?
Re: ID problems in 3.0
Posted: Tue May 28, 2013 2:18 pm
by aymanh
Hi James,
There was no bug there, based on earlier messages in the trail it's either of:
- Mass properties unspecified (Mass, Inertias)
- Infinite or very large forces being applied (because of poorly initialized muscles in one case)
If masses and inertia look good then I'd try to zone in on which forces are causing the problem by taking suspect forces out then adding them one at a time to pinpoint the problem.
Hope this helps,
-Ayman