Inverse Dynamics error

[Max Rusch]

Hi
I'm using the model of the upper limb.
Realized the steps of scaling and inverse kinematics.

Now, I'm in the step of inverse dynamics. However, I have difficulty because the examples in the manual and the examples files are in gait.

I'm having this error when I place the input file given by IK (mot or attic) in the GUI inverse dynamics:

Tool execution failed. Check messages window for details.

In the message window appears this:

Exception:
InverseDynamics: ERROR-mass matrix is singular
file = .. \ .. \ .. \ OpenSimBranch \ OpenSim \ Analyses \ InverseDynamics.cpp
line = 423

Can anyone help me with this problem?

Thank you for your attention,

Regards, Max Rusch

[Ayman Habib]

Hi Max,

The Upper extremity model may not have mass properties assigned (Mass and Inertia's should be non-zero so that you don't get infinite acceleration due to a very small force).

Hope this explains,
-Ayman

[Max Rusch]

Thank you very much!!

However, in the model of upper limb, I have some phantom body:
clavphant (between clavicle and scapula); scapphant, humphant and humphant1 (between scapula and humerus). These phantom bodies are also without mass, center of mass and inertia. The question is: I need to complete these values for these phantom segments or not?
Also, do not understand the function of these phantom bodies.
Although it seems that each one allows a certain movement in a single plane, and the combination allows three planes of motion for the humerus. Correct?

Thank you for your attention.

Sincerely, Max Rusch

[Karinne Ramirez]

Hi,
I also have the same error and I also realize that the upper body has mass, center of mass and inertias set to zero! I'm analyzing not just the upper body but the whole body. My question is: is there a table or a formula to get those data? I'm new in this biomechanical field and I will really appreciate your help . I'm reading the book of "Biomechanics and motor control of human movement" of David Winter, and I found a Table that has the segment weight, is it possible to use that?
Thanks for your help

Karinne

[Ayman Habib]

Hi Karinne,

I'd suggest you contact the model authors since they would be more familiar with the assumptions in their model, in particular the different segments and joints. Also check if the inertia reported is wrt center of mass (com) or geometry origin, where's the com, etc.

Good luck,
-Ayman

[Max Rusch]

Hi Karinne,

The book "Biomechanics and motor control of human movement" of David Winter and the book "Kinetics of human motion" of Zatziorsky have the information of mass, center of mass and moment of inertia. However, in these books the information of anthropometric inertial parameters is to segments how arm and forearm; and in some models of opensim you need inform parameters for each bone (eg, ulna, radius).
My suggestion is to measure the bone volume and get data of bone density in other books and obtain these parameters by mathematical equations.

Hope that helps.

Regards, Max

[Karinne Ramirez]

ok. Thanks for the suggestion

[Wayne Johnson]

Was the "phantom" mystery ever resolved? I am now trying to use the MoBL-ARMs UE (https://simtk.org/projects/upexdyn) but do not understand what the scapphant, clavphant and humphant bodies are for especially since they all have zero inertial properties.

[Woojae Kim]

I would also like to know about the "phantom bodies". In the model that I use (MoBL-ARMS dynamic upper limb model - OpenSim 3.2+), they do have some small mass assigned to it (10^-4 kg). Would mass in these bodies change inverse kinematics and dynamics results?