I saw the following message after a failure RRA.
CMC.computeControls: WARN- The optimizer could not find a solution at time = 0.625000.
If using the fast target, try using the slow target.
Starting at a slightly different initial time may also help.
Does anyone know why it helps by changing the starting time.
Does it mean that the intropolation algorithm is not stable?
Why does the different initial time help
RE: Why does the different initial time help
Hi Kang, have been watching you for a long time period, I think a lot of us in this forum encountered the same problem.
Have you passed that time point? What I meant is that when I change the different starting point, I still could not continue the CMC. Do you think I should start from the previous steps? Like Scaling or IK?
How do you think a perfect scaling or IK would help to reduce the residual or helpful to CMC?(or RRA)
Thanks in advance. Really spending a lot of time of OS but still felt so much frustrating about myself.
Yi.
Have you passed that time point? What I meant is that when I change the different starting point, I still could not continue the CMC. Do you think I should start from the previous steps? Like Scaling or IK?
How do you think a perfect scaling or IK would help to reduce the residual or helpful to CMC?(or RRA)
Thanks in advance. Really spending a lot of time of OS but still felt so much frustrating about myself.
Yi.
RE: Why does the different initial time help
No. I did not resolve the problem.
I am trying to reproduce the successful stories from this forum and comparing their settings with mine.
I am trying to reproduce the successful stories from this forum and comparing their settings with mine.
-
Rebecca Routson
- Posts: 25
- Joined: Thu Jul 30, 2009 10:19 am
RE: Why does the different initial time help
Different initial time could help for a number of reasons.
1. Sometimes at your initial start time your kinematic data starts with all of your marker placements at zero and that is why you would have a person-ball.
2. You should only be running RRA and CMC while you have forceplate data. If your subject starts or ends your trials off of your forceplates you should clip your trials to only the times where there are forces. If you don't do this you will tend to have your subject either fly off the ground or sink into it. You can plot your residual forces either after RRA or ID to check where this is happening or just look at your forceplate data.
3. Various other reasons I just cant think of right now.
Try changing your start time significantly and if you still cannot run your simulation, that probably is not your problem. Your error message also refers you to changing to the slow target.
Hope that helps.
Becca
1. Sometimes at your initial start time your kinematic data starts with all of your marker placements at zero and that is why you would have a person-ball.
2. You should only be running RRA and CMC while you have forceplate data. If your subject starts or ends your trials off of your forceplates you should clip your trials to only the times where there are forces. If you don't do this you will tend to have your subject either fly off the ground or sink into it. You can plot your residual forces either after RRA or ID to check where this is happening or just look at your forceplate data.
3. Various other reasons I just cant think of right now.
Try changing your start time significantly and if you still cannot run your simulation, that probably is not your problem. Your error message also refers you to changing to the slow target.
Hope that helps.
Becca