Inverse Dynamics: Generalized Forces not being output

[Mathew Varre]

Hello All,

I have been using inverse dynamics tool to compute the generalized forces and torques in an upper extremity model (Arm2013). I have input my reaction forces (N), Point of Force Application (m) and moments (N-m) into the ID setup file. The output .sto file has only moments at the shoulder, elbow and wrist joint and forces only at the thorax. I do not see forces at the joints (shoulder, elbow and wrist) in this output file. I checked some of the output files and setting files for ID in the lower extremity models and they don't have all of the joint forces as well. I am just wondering how to output all the joint forces in the .sto file when I run ID.

Thank you for your help.

Regards
Mathew

[jimmy d]

Hi Mathew,

Generalized forces correspond to generalized coordinates. ID will output the generalized forces for each generalized coordinate available in the model. If you have a three-rotation joint (a ball-joint for example) you will get a total of three generalized forces (one for each rotation coordinate)

By your description you are after the contact loads. You will need to run a joint reaction analysis to do this.

Cheers,
-james

[Mathew Varre]

Hi James,

Thank you for your reply. I agree with your statement that ID gives generalized force at each generalized coordinate available in the model. But when I run ID, the .STO file output contains only the moments for these generalized coordinates but NOT the forces. I tried different .MOT headers and used an already available file from other models in OpenSim package but was not successful in getting the generalized forces outputted.

Does ID output these at all? I could find only pelvic (x/y/z) forces (in my case thorax) but not knee and ankle joint forces in the . STO file for gait2354 and gait2392 models in the package. (As show in the picture attached).

If you have any suggestions and/or troubleshooting ideas to output these generalized forces for all the coordinates in ID, I would really appreciate it.

Regards
Mathew

[jimmy d]

Hi Mathew,

ID calculates the forces about a coordinate that are required to generate the motion of that coordinate. If there is no coordinate, there is no motion and there is no force to be generated. If you have a knee joint that is modeled as a hinge, there is only a single coordinate and only one (rotational) generalized force. It is a 1:1 relationship between the number of generalized coordinates in the model and the number of generalized forces output from ID.

From the .sto picture you attached, the pelvis has 3 rotational coordinates (list, tilt, rotation) and 3 translational coordinates (tx, ty, tz). Each coordinate then has its corresponding generalized force (moments for rotational and force for translational). The hip, knee and ankle are modeled with only rotational coordinates and the generalized forces of these rotational coordinates are output as 'moments'. There are no translational coordinates at the hip, knee and ankle so there will not be any translational forces output.

If you want the forces exerted between the tibia and the femur, you are describing Joint Reaction Analysis. Joint contact is a different concept than ID.

There is a good description of the differences between ID and Joint Reaction here; http://simtk-confluence.stanford.edu:80 ... s+Analysis

There is also webinar done by Matt DeMers on calculating joint contact; https://www.youtube.com/watch?v=ly4rhOOifO0

Hope that helps,
-james

[Mathew Varre]

Hi James,

Thank you very much for your prompt and detailed response about the ID. This is very helpful to me. I do understand now that 'rotational coordinates' will provide 'rotational moments' at the joints and 'translational coordinates' would provide 'translational forces'. I was thinking I could get the reaction forces at the joints and moments using ID tool, like I did in a typical ID models which I have used in the past. But your post answers that for me that I would need to define 'translational coordinates' in the model to output forces at a joint, which I clearly did not think of until now. I did not confuse ID with joint reaction analyses as I am aware that I would need muscle forces/motors/actuators and run FD to get the joint contact loads. But the information you have provided would be helpful when I reach that part of simulations.

Thank you once again for your reply.

Regards
Mathew