Dear all
I am running static optimization (SO) and it fails for some of my subjects. I am trying to figure out if it comes from the inverse dynamic process or if the problem is in that my model is weak and so on.
I have plotted the output of inverse dynamics (ID) and it concerns me since something seems amiss. I see spikes at some interval of time. I am a bit at a loss on why this might be happening so any suggestions on how to go about it are very much welcomed!
I have tried different trials and other subjects and this issue seems consistent although some cases are less severe and SO seem to work. Also my inverse kinematics seems fine. Any ideas/suggestions what might be happening with my ID step?
I have attached 3 plots from 3 different subjects and the SO is unsuccessful for "Inverse dynamics S3_SO failed" (first pic shown).
Thanks for any help
Inverse dynamics problem
Re: Inverse dynamics problem
Hi Karelia-
It it may be an issue with the GRF data your inputting into the model. Ill PM you and we can discuss.
-James
It it may be an issue with the GRF data your inputting into the model. Ill PM you and we can discuss.
-James
Re: Inverse dynamics problem
Karelia-
It seems that the GRF processing may be causing the issue. When I run ID without the external loads the joint moments seem to be giving nice values. This, along with no apparent issue with the model kinematics, I would suggest playing around with your settings for GRF processing to see if that effects the ID results.
Also a note, I would track thorax movement (if you had experimental markers) locking the lumbar could cause alot of issues with RRA and static optimization. The first thing that is moved to reduce residuals is the HAT segment, as it has the largest mass. If the position of the HAT is incorrect the muscles may get overloaded during static optimization and cause a crash.
Please let me know if you get improved results,
-james
It seems that the GRF processing may be causing the issue. When I run ID without the external loads the joint moments seem to be giving nice values. This, along with no apparent issue with the model kinematics, I would suggest playing around with your settings for GRF processing to see if that effects the ID results.
Also a note, I would track thorax movement (if you had experimental markers) locking the lumbar could cause alot of issues with RRA and static optimization. The first thing that is moved to reduce residuals is the HAT segment, as it has the largest mass. If the position of the HAT is incorrect the muscles may get overloaded during static optimization and cause a crash.
Please let me know if you get improved results,
-james
-
Karelia Tecante
- Posts: 7
- Joined: Thu Jun 14, 2012 10:35 am
Re: Inverse dynamics problem
Dear James,
I looked at the grf file and it seems ok. I plotted the grf and hip flexion and it looks fine to me (see fig). I believe the SO is having problems at heel strike on both FPs (biggest peaks in the graph). I added actuators for every joint and it seems to be running better. I believe I am having SO issues for subjects who have a strong heel strike since SO fails exactly at the highest peak time which is the heels strike time.
Kind regards
I looked at the grf file and it seems ok. I plotted the grf and hip flexion and it looks fine to me (see fig). I believe the SO is having problems at heel strike on both FPs (biggest peaks in the graph). I added actuators for every joint and it seems to be running better. I believe I am having SO issues for subjects who have a strong heel strike since SO fails exactly at the highest peak time which is the heels strike time.
Kind regards
- Attachments
-