Hello
This trail failed says Orientation close to Gimbal lock in the warning, The Warning shows many lines of “Constraint no. 519(ect) above error tolerance 0.000001”. But then there was an Error Message.
You can see the warning and error message in the attachments.
Can you please take a look at give some recommendation how to fix this type problem please?
What does this mean, “more appropriate settings of the kinematic solver” please?
Sincerely thanks,
Damon
Hi Damon,
This is pretty much the same problem as the previous one. You need to figure out which kinematic constraints cannot be satisfied. Have a close look at the model and see what behaves abnormal.
Kind regards,
Pavel
Hi Pavel
Wow, it seems like looking for needle in haystack.
So far these methods are known to me.
1- Improve the initial positions.
2- figure out which kinematic constraints cannot be satisfied. Have a close look at the model and see what behaves abnormal.
a. Can look for markers that are not close to each other.
Q-1
Would the previous Warning message (attached) be a clue where to look? See the attachment about singular position, maybe this is the problem?
Q-2
Do you have any other suggestions for methods in the search for abnormality in the model?
Q- 3
Is it possible to increase the error tolerance above 0.000001 ? See Attachment 2 please.
Q-4
The Error message says, "more appropriate settings of the kinematic solver may allow a solution. Does this mean there is some setting for kinematic solver?
Sincerely hope your having great time,
Damon
Hi Damon,
Yes, you should go with #2 in your list. Some of the constraints are conflicting in your case. For example, box markers “pull” the box anteriorly, whereas connection to the hands “pulls” them backwards - this results in unnaturally large error in one or another constraint.
A1: Not necessarily, but unlikely in this case.
A2: I may have some, but i need to see the exact model you are working on (but I would prefer to give some guidance so you can find the problem yourself). And as I understood we made your previous trial work and now looking at something else.
A3: Yes, this is one of the settings for the kinematic solver. But please note that by increasing error tolerance you allow inaccuracies in your model.
A4: Yes, please look at the reference manual and read more about AnyBodyStudy class (in particular more about members InitialConditions and Kinematics).
Kind regards,
Pavel
Hi Pavel
I moved the box very close to the initial position location, I also moved all the bones much closer to the initial position location, but there is still the same ERROR problem. There has been over 200 trails which ran ok with the box and bones not as close, so I am thinking the initial positions may not be the issue causing the ERROR.
So far the only one that we got to work was the foot one.
[FONT=Calibri][SIZE=3]Are these the lessons to review?[/SIZE][/FONT]
Lesson2: Initial Conditions
This was good to read.
“Instead of simply running the InitialConditions operation, you can also single-step through it to see what it does. This is done by clicking the “Step” button instead of the “Run” button.
The first step re-establishes the load-time conditions. This means that it positions the modes as it was when you loaded it. This is useful for identification of kinematic problems.”
http://www.anybodytech.com/fileadmin/AnyBody/Docs/Tutorials/chap4_A_study_of_studies/lesson2.html
Lesson 3: Kinematic Analysis
This does not give a method to fix the problem, only describes the problem. “Actually, even when you have a kinematically indeterminate system, the Kinematics operation can fail. This is actually very easy to picture. Sometimes the segments of the model may be configured such that they cannot reach each other, or in such a way that they interlock. The real world is full of that sort of mechanisms: Car doors that get stuck or refuse to close, locks that will not unlock, or stacked glasses that wedge inseparably into each other. Computer systems that model the real world will have them too, and just like the real world it can sometimes be difficult to find out what the problem is.”
http://www.anybodytech.com/fileadmin/AnyBody/Docs/Tutorials/chap4_A_study_of_studies/lesson3.html
Pavel
I’ve attached the Vicon motion capture, could you please take a look at it, maybe try running it and see if you can get the same ERROR?
Sincerely thank you very much.
Damon
Hi Damon,
Firstly I have not studied the c3d file in detail…
If the model fails in the first frame it can be two things:
*Initial position: this is needed to the get the model started in a location close enough to the actual solution.
*Kinematic constraints problems:
Please review if the markers are all good in the trials causing the problem… so check there are no dropouts.
If the model contains constraints to some environment objects like a box, please try to simplify to find the root of the problem. So if possible disable the box, to have model only driven by markers.
Once you have done this and made sure it runs, add the box back into the model, do this stepwise. So initially drive all six dof of the box using some resonable driver you will make up virtually. Then remove one driver DOF at the time from the box and replace it with a constraint between hand and box.
This approch will possibly give you an idea on which DOF are causing the problem. I have no idea if you have a rotational constraint between hand and box, but if so i would recommend to replace it by a combination of linear constraints, this can sometimes make the problem easier to solve.
Hope it mades sense
Best regards
Søren
Hello Søren
The model does not fail in the first frame. This trail fails at the 51 frame out of 221 frames.
The model with box has ran ok with many trails, over 200. So I think the model with box is ok.
I am not sure what is causing this failure. Ive checked the markers many times, and set the initial position very close. I’ve attached the model, hopefully you can take a closer look at it please.?
Also, would AnyBody wish to publish this model in the repository? If so id need to clean up some of the code notes, and its ready.
There is also a validation publication for this model. http://www.sciencedirect.com/science/article/pii/S0169814115300652
Sincerely thank you very much,
Damon
Hi Damon,
I have tried to load the model without any changes… so far it has passed step 121 and is running .
Please advice on how you made it fail…
I am running AMS 6.1.1 .
I am running the inverse dynamic study, this is how the model was uploaded… if this should have been kinematics… then i do not understand why the optimized motions is already ok…
Best regards
Søren
Hello Søren
Thank you for running the model. When I ran it the kinematics (bones) ran fine, and when I ran the dynamics (muscles) it failed at step 51. I tried it several times and it failed each time. This is weird that it is running ok now.
Did it run complete without failing?
Possibly I need to use AMS 6.1.1, what does this mean please?
Sincerely thank you very much,
Damon
Hi Damon,
I have checked the model ran all the way through…i am using AMS 6.1.1.4986 64 bit version.
If you have valid maintenance you should be able to download and run this version.
AMS 6.1.1 contains various improvements in the kinematic engine which could possibly explain why it runs in this version and not in yours? which one are you using?
Best regards
Søren
Hello Søren
We uploaded the 6.1.1 software and it is running so much better. Six of 7 trails that failed because of the Gimbal lock problem have completely ran.
The trail I sent you also ran past the 51 step (previous fail point), but it failed at 210 of 221 steps. The error pointed to a muscle, but when looking at that muscle it looks fine. Not sure what to do. Did you have any problems at step 210 when you ran it?
Sincerely thank you very much!
Damon
Hi Damon,
I am happy that the new version is running better.
I reran the model and it runs all the way to the end, so i am not sure if you have modified the model a bit?
What is the exact error you get?
Best regards
Søren
Hello Søren
Thank you, these were the last few trails needed to complete the data for the 240 trails.
I hadn’t changed anything in the model. The model you have is the same as the one I have. But its possible, that your muscle file is different, since I didn’t send you the body files I am using.
I attached the screen shot showing the error. Hope this helps us to figure out the problem. The error points to a muscle.
Sincerity thank you very much
Damon
Hi Damon,
I think i hooked the model up to run on the default AMMR1.6.6, maybee this exaplains the different behaviour.
Best regards
Søren
Hello Søren
Thank you, I’m sure this is the reason. Can I download the 6.6 software?, to test it sometime. If so can you send me the link to where it is please.
Sincerely thank you
Damon