Elbow disconnected

Hello all,

I use an upper extremity marker set to drive my model of the trunk and right arm. I proceed to use the MotionAndParameterOptimizationModel and the InverseDynamicModel to calculate Joint Loading. When I switch the ‘OptimizeAnthropometricsOnOff’ on for the UpperArm and LowerArm the Elbow joint is disconnected after the Optimization. Because I do not know each segments length exactly, I need the Optimization to complete my movement (maximal reaching task). Otherwise the motion is not followed completely and the arm is not fully stretched or goes into gimbal lock.

How can I fixate the connection between the upper and lower arm during optimization?

Greets,

Hans

Sorry, Hans.

I have just noticed the private message. Do you remove the elbow joint for the optimization sequence? Why has it disconnected?

But the joint is somehow missing - you can add back by using the revolute joint, for example.

Could you explain the problem once again?

Regards,
Pavel

I do not know if the Elbow joint driver is removed or where to find it exactly. It is not present in the ExtraDrivers.any.

I use the optimization to alter the lengths of the upper- and forearm. After I run the MotionAndParameterOptimizationModel the optimal motion has a disconnected elbow.

I have used an older version of AnyBody and never experienced this problem. I’ve recentely switched to V6 and encountered several errors with my older Model. I’ve adjusted some functions according to the update, but did not touch any drivers.

You can doubleclick on your study that is optimization to see what kind of constraints are there, and which segments are involved. It should give you an overview of what is going with those bones. Somehow the elbow joint is missing - i am not sure why.

P.

I’m processing an optimization right now, what should I look for exactly?

I’ve noticed an error in which the humerus is rotated 180 degrees. The Elbow C3D marker is shown on the medial side of the body and the humerus is rotated so it fits with the AnyBody marker. However the C3D marker is placed on the lateral epicondyle, so I don’t know why the humerus is twisted. Should I alter the initial segment lengths?

The forearm is in the correct direction, but it is sticking through the humerus.

Hans,

it is actually a somewhat difficult problem. The motion optimization study performs an overdetermined kinematic analysis, which is able to violate joint constraints. It is an unlucky scenario.

Typically the solution would be to add some soft constraints to limit the rotation of the bones. You could check an example of how the default drivers are constructed in the StandingHuman model. In other words, you need to help the system to solve your optimization problem by adding some extra constraints.

Regards,
Pavel

I’m afraid I do not have the necessary skills yet to do this in a short time. I’ve registrered for a course later in March, but this will be too late. Is there any way I can upload my model so you or a colleague can have a look at it?

You could try uploading to the Debug section, we will look at it when we can, but unfortunately i cannot promise a very quick solution - we are quite busy too.

Kind regards,
Pavel

Allright, thanks for the help.
I’ve added my model in http://forum.anyscript.org/showthread.php?t=3825

Greets,

Hans