GRF in 1.9.1 and 2.0.2

Provide easy-to-use, extensible software for modeling, simulating, controlling, and analyzing the neuromusculoskeletal system.
POST REPLY
User avatar
Kang Li
Posts: 59
Joined: Tue Dec 11, 2007 8:36 am

GRF in 1.9.1 and 2.0.2

Post by Kang Li » Mon May 10, 2010 9:29 am

Hi All

I know the method for applying GRF to the feet has been improved in 2.0.2. It is more flexible.
However, if I want to implement the way I applied the GRF in 1.9.1, what would be the parameters I should use in 2.0.2. Should I choose Point Force or Body Force, should I treat the Force global or not? Can anyone kindly provide some guidances?

Thanks

User avatar
Ayman Habib
Posts: 2235
Joined: Fri Apr 01, 2005 12:24 pm

RE: GRF in 1.9.1 and 2.0.2

Post by Ayman Habib » Mon May 10, 2010 12:49 pm

Kang,

There's auto conversion code built into the various Tools so that if a xxx_grf.mot file was used to specify external forces in an earlier version of the setup files, a new xxx_grf.xml file is created and it contains 2 prescribed forces that have the following attributes:
- Point Force: Force Columns(vx,vy,vz)
Point Columns (px, py,pz)
- Torque Columns (torque_x,y,z)
Applies torque (true), Force is global (true), Point is global (false)

You should be able to create these from the GUI now, though it's recommended that you rename column labels to be unique first to avoid confusion.

The force-is-global/point-is-global flags are a bit misleading because the tools do assume that you have a kinematics-for-external-loads-file specified and do transform the "point of application" to the frame of the body where the force is applied based on those kinematics.

Hope this explains, and please let us know if you have questions,

-Ayman

User avatar
Kang Li
Posts: 59
Joined: Tue Dec 11, 2007 8:36 am

RE: GRF in 1.9.1 and 2.0.2

Post by Kang Li » Wed May 12, 2010 12:14 pm

Hi Ayman,

I tried difference GRF implmentations. For the same movement, the Static Optimization works well in 1.9.1 while in 2.0.2 the Static Optimization works if the forces are specified as body forces . The SO could fail if the GRFs are specified as point forces. Does that mean the conversion is different from what you describe above?

Thanks,
Kang

User avatar
Ayman Habib
Posts: 2235
Joined: Fri Apr 01, 2005 12:24 pm

RE: GRF in 1.9.1 and 2.0.2

Post by Ayman Habib » Wed May 12, 2010 12:27 pm

Kang,

The description I gave you is just code reorganization and shouldn't affect your results or ability to run StaticOptimization, so probably there's something else that's causing the difference in behavior. I'd suggest you file a bug report and include all the files necessary to reproduce this in both 1.9.1 and 2.0.2 and then we can investigate.

One issue that may have tripped you is the auto-migration/renaming for Force/Actuator classes between the versions. The code to perform this migration keys off the DocumentVersion tag at the top of the xml files so if you manually copy and paste between xml/osim files without paying attention to which names/tags go with which version you can easily get in trouble.

-Ayman

POST REPLY