Problem with Marker optimsation

Main Forum
1

Hi,

I've been trying to run the ParameterOptimisation in order to optimize marker positions (using Standard Plug-in Gait) based on Plug-in-gait_Simple/FullBody model.

I've performed an optimisation regarding a gait task previously, which has been completed enlarging kinematicTolerance to 10^-4, since with lower tolerance value, kinematic analysis fails.

Later, I've performed an other parameter optimisation based on grabbing a drill from a table and drilling. In this operation I've deactivated segment length optimisation, instead optimal dimensions previously computed have been inserted because he was the same subject as before.
I've tried to perform marker position optimisation as I did before for the gait task but it failed.

I get the following warning messages every time that I load the model. May I have to worry about them?

WARNING(OBJ1) :   MarkerProtocol_Opt-Upper.any(64)  :   MarkerWeightsFun  :  Some of the detected transitions in the data set were skipped (i.e. not entered into 'Ts') because they too close compared to the transition time 'dT'. 54 transitions were skipped!
WARNING(OBJ1) :   MarkerProtocol_Opt-Upper.any(91)  :   MarkerWeightsFun  :  Some of the detected transitions in the data set were skipped (i.e. not entered into 'Ts') because they too close compared to the transition time 'dT'. 1 transitions were skipped!
WARNING(OBJ1) :   MarkerProtocol_Opt-Upper.any(225)  :   MarkerWeightsFun  :  Some of the detected transitions in the data set were skipped (i.e. not entered into 'Ts') because they too close compared to the transition time 'dT'. 1 transitions were skipped!
WARNING(OBJ1) :   MarkerProtocol_Opt-Upper.any(542)  :   MarkerWeightsFun  :  Some of the detected transitions in the data set were skipped (i.e. not entered into 'Ts') because they too close compared to the transition time 'dT'. 1 transitions were skipped!
WARNING(OBJ1) :   MarkerProtocol_Opt-Upper.any(568)  :   MarkerWeightsFun  :  Some of the detected transitions in the data set were skipped (i.e. not entered into 'Ts') because they too close compared to the transition time 'dT'. 64 transitions were skipped!

Then, I run some Parameter Optimization changing which markers being optimized and I get the following errors:

Failed to resolve kinematic constraints. Newton relaxation too small. (final kin. error = 3.612224E+00)Constraint no. 143824 above error tolerance 0.000100, error = 0.1474463804458.Constraint no. 143843 above error tolerance 0.000100, error = 0.6876287137966.

.Previous AbsSumKinError before decrease of step length: 55.748214Failed to resolve kinematic constraints. Newton relaxation too small. (final kin. error = 1.134530E+01)Constraint no. 143824 above error tolerance 0.000100, error = 0.1473977581444.Constraint no. 143851 above error tolerance 0.000100, error = 0.0416170747519..

I would really appreciate if someone could explain me what these errors/warnings messages stand for, and how I can fix them.

Thank you in advance to anyone that would like to help me!

2

Hi @Morriszs & welcome to the forum!

The MarkerWeightsFun warning you are getting happens when the marker have drop outs in the data. When that happens multiple times and very close to each other there is not enough time for the weight function to transition and it will treat the whole period as one long drop out. This should not be big concern for you.

The failing of the parameter identification however is of course bad.
I think you need to check that all the model markers are located where to expect them to be, some marker positions can be affected by adipose tissue and other sources of error.

You can use the search function here on the forum to find related posts and see what others have tried and you can use our wiki page for some helpful tricks.

Best regards,
Bjørn
AnyBody Technology

3

This topic was automatically closed 125 days after the last reply. New replies are no longer allowed.