I have a problem related to the scapular position when running MarkerTracking using the MoCap full body model (Multi trial). AMMR v 2.2.3 and AMS v 7.3. I have attached a picture showing the problem and I get the following error message:
1.0.1.0.0.2) ...Kinematic analysis completed.
1.0.1.0.0.3) ...Dependent variables are fully updated.
1.0.1.0) Kinematic analysis...
Progressing to solve kinematic optimality conditions and hard constraints.
Progressing to solve kinematic optimality conditions and hard constraints.
Progressing to solve kinematic optimality conditions and hard constraints.
Progressing to solve kinematic optimality conditions and hard constraints.
WARNING(OBJ.MCH.KIN6) : Interface.any(260) : GHMeasure : Close to singular position : Orientation close to Gimbal Lock, i.e., first and third axis of rotation being parallel
Failed to solve kinematic optimality conditions and hard constraints after 5 fallback attemps.
Constraint violations for study 'Main.Studies.MarkerTracking' :
Constraint #18 is above tolerance 1e-06, error = 0.000001, constr. #0 in 'Main.HumanModel.BodyModel.Trunk.JointsLumbar.T12L1Jnt.Constraints'.
.
.
.
Constraint #423 is above tolerance 1e-06, error = 0.000181, segment constr. 'Main.HumanModel.BodyModel.Left.ShoulderArm.Mus.DeltoidWrappingAnterior.WrappingSegment'.
ERROR(OBJ.MCH.KIN3) : AnyMocapModel.any(41) : MarkerTracking.Kinematics : Kinematic analysis failed in time step 67 : Position analysis is not completed
I have played around with settings in the ExtraDrivers.any file but without success. The drivers (that I think are related) that I use now are:
#define BM_MANNEQUIN_DRIVER_STERNO_CLAVICULAR_LEFT ON #define BM_MANNEQUIN_DRIVER_STERNO_CLAVICULAR_RIGHT ON #define BM_ARM_SHOULDER_RHYTHM RHYTHM_SOFT
I have tried activating soft rhythm for the arm-clavicula rotation and also turned the arm_deltoid_wrapping off but without success.
Any advice on how I should proceed the debugging is more than welcome!
I can see in the error message that there is a gimbal lock in the GH joint. This is not a problem unless this joint has a driver applied on it. If you have soft drivers on the GH joint that would add a driver on this measure, so if this is the case please disable these soft drivers.
Please also try to disable the rhythm, it is possible that the rhythm is using the GH measure which has gimbal lock.
Does this error happens in the first frame or later in the analysis?
If it is the first frame you can try to improve the initial conditions a bit that might
The most recent AMMR 2.3 version has some improvements related to the shoulder kinematics is it possible that these may improve your model.
If this does not help we could take a look at the model, if possible.
I did not have any of the glenohumeral drivers on so the problem must be somewhere else. I disabled the soft shoulder rhythm and then the model ran passed the point where it failed before (at step 67 so not in the beginning). However, it failed later in step 157 and the scapula control still seems to have some problems since it is instead penetrating the ribs. I have a feeling that this problem also causes an artificial twist of the spine, which is really something I would like to avoid as I will look at forces in the lumbar spine and lower-back muscles.
I tried to attach the model to this post but it was too big. Is there another way I can upload or send you the model?
Would I be able to switch to AMMR 2.3 just by placing my folders (corresponding to the ones I uploaded) to my own copy of the AMMR 2.3 folder instead or would I encounter some compatibility problem with my files? I'm asking since I had troubles before, but that was between version 1.6 something and version 2.x if i remember correctly.
Going from 2.2.3 to 2.3 should not be a problem it should be compatible.
AMMR 2.3 would also allow you to use new muscle wrapping which is significantly faster.
