SimTK: OpenSim: [#3031] OpenSim-generated du/dt curves appear to be incorrect

Share
Follow	Project Follow publicly Follow privately Unfollow See followers (304) List Map More info

Trackers

Date:
2021-01-02 06:35

Priority:
3

State:
Open

Submitted by:
B.J. Fregly (fregly)

Assigned to:
Nobody (None)

Summary:
OpenSim-generated du/dt curves appear to be incorrect

Detailed description
I just ran the Analyze tool, including Kinematics and Body Kinematics, on an OpenSim model for which I had just generated Inverse Kinematics results. Since I cannot read in my own u and du/dt values, I wanted to verify that when OpenSim differentiated the q values in my IK-generated .mot file processed with a 6 Hz low pass cutoff frequency, it generated the correct u and du/dt curves. I need all three types of curves (q, u, and du/dt) for an IMU motion measurement project on which I am working. Unzip the attached file and run the included Matlab program, which reads in the q, u, and du/dt data output by the Analyze Tool, fits the q curves with a smoothing Bspline, and then analytically calculates the associated u and du/dt curves for comparison. I included an analytical test function at the start to confirm that the Bspline function is creating the correct u and du/dt curves. As you will see, when du/dt curves output by the OpenSim Analyze Tool was WAY to large and not the correct shape either. So unless I am missing something (which is always possible), it appears that OpenSim is not calculating the du/dt curves correctly. I don’t know if that means Body Kinematics and Point Kinematics results will be wrong as well, but I definitely need Body Kinematics to work properly for the project on which I am working. Let me know if you have any questions about this bug report. Hopefully I am just doing something silly!
I just ran the Analyze tool, including Kinematics and Body Kinematics, on an OpenSim model for which I had just generated Inverse Kinematics results. Since I cannot read in my own u and du/dt values, I wanted to verify that when OpenSim differentiated the q values in my IK-generated .mot file processed with a 6 Hz low pass cutoff frequency, it generated the correct u and du/dt curves. I need all three types of curves (q, u, and du/dt) for an IMU motion measurement project on which I am working. Unzip the attached file and run the included Matlab program, which reads in the q, u, and du/dt data output by the Analyze Tool, fits the q curves with a smoothing Bspline, and then analytically calculates the associated u and du/dt curves for comparison. I included an analytical test function at the start to confirm that the Bspline function is creating the correct u and du/dt curves. As you will see, when du/dt curves output by the OpenSim Analyze Tool was WAY to large and not the correct shape either. So unless I am missing something (which is always possible), it appears that OpenSim is not calculating the du/dt curves correctly. I don’t know if that means Body Kinematics and Point Kinematics results will be wrong as well, but I definitely need Body Kinematics to work properly for the project on which I am working. Let me know if you have any questions about this bug report. Hopefully I am just doing something silly!

Comments (1)
Attachments
Changes

Add A Comment:

Message
Date: 2021-01-05 08:15 Sender: B.J. Fregly I just did an Autolev-based calculation of Body Kinematics results from the Analyze Tool, and while the positions and velocities were spot on, the accelerations from OpenSim were again WAY to large compared to the accelerations calculated by Autolev. So it appears that the du/dt values are not just being output incorrectly - they appear to be incorrect internally as well. To do this test, I created a lower body OpenSim model with IMUs on it that matches experimental measurements. I then created an identical Autolev model that had exactly the same joint kinematic structure and IMU positions/orientations on the body segments. Given that the positions and velocities match perfectly, I am convinced that the OpenSim acceleration calculations are incorrect.

Attachments:

Size	Name	Date	By	Download
377 KiB	OpenSim Bug1.zip	2021-01-02 06:35	fregly	OpenSim Bug1.zip

Field	Old Value	Date	By
File Added	740: OpenSim Bug1.zip	2021-01-02 06:35	fregly

SimTK is maintained through Grant R01GM124443 01A1 from the National Institutes of Health (NIH). It was initially developed as part of the Simbios project funded by the NIH as part of the NIH Roadmap for Medical Research, Grant U54 GM072970.

Version 4.2.2. Website design by Viewfarm. Icons created by SimTK team using art by GraphBerry from www.flaticon.com under a CC BY 3.0 license. Forked from FusionForge 6.0.5.

Feedback