Hi all,
I am using the inverse kinematics tool on a number of motion trials I have collected, and two are failing part way through the IK calculations. In one trial, the subject just walks back and forth along a straight line for about 23 seconds. The other trial is the same except the subject is jogging.
About 14 seconds into the walking trial, the IK tool fails with this error:
AssemblySolver::track() attempt Failed: SimTK Exception thrown at Assembler.cpp:896:
Method Assembler::track() failed because:
Optimizer failed with message: SimTK Exception thrown at InteriorPointOptimizer.cpp:261:
Optimizer failed: Ipopt: Maximum iterations exceeded (status -1)
Assembly error tolerance achieved: 0.188241544167549 required: 1e-006.
InverseKinematicsTool Failed: AssemblySolver::track() attempt failed.
I get the same error about 4 seconds into the jogging trial.
Can someone please explain to me what causes these errors and suggest some way to remedy it?
Thanks,
Anthony
Inverse Kinematics fails part way through
- Anthony Ligouri
- Posts: 13
- Joined: Fri Feb 24, 2012 11:47 am
- Ayman Habib
- Posts: 2255
- Joined: Fri Apr 01, 2005 12:24 pm
Re: Inverse Kinematics fails part way through
Hi Anthony,
The error reported indicates that OpenSim couldn't solve the assembly problem for the specific frame (starting from a solution to the previous frame). Possible reasons:
- Sudden or big change in kinematics
- Constraints in the model being violated in this frame (while they were not in previous ones).
- Dropped or mislabeled markers around this frame (that would disrupt the continuity of the solution)
I'd suggest you inspect your data and see which of these cases apply. If none is obvious please file a bug report and attach the model and all necessary files and we'll investigate.
Best regards,
-Ayman
The error reported indicates that OpenSim couldn't solve the assembly problem for the specific frame (starting from a solution to the previous frame). Possible reasons:
- Sudden or big change in kinematics
- Constraints in the model being violated in this frame (while they were not in previous ones).
- Dropped or mislabeled markers around this frame (that would disrupt the continuity of the solution)
I'd suggest you inspect your data and see which of these cases apply. If none is obvious please file a bug report and attach the model and all necessary files and we'll investigate.
Best regards,
-Ayman