Reserve actuators optimal force value

Provide easy-to-use, extensible software for modeling, simulating, controlling, and analyzing the neuromusculoskeletal system.
POST REPLY
User avatar
Charles David Sasportes
Posts: 15
Joined: Thu Jan 23, 2020 6:36 am

Reserve actuators optimal force value

Post by Charles David Sasportes » Sun Apr 19, 2020 2:55 pm

Hello everyone,

I have a pretty basic question about reserve actuators.

If I understood correctly, reserve actuators are added to a model when running simulations, for example, static optimization, in order to "help" the model achieve the desired motion even if the muscles are too weak to do it by themselves. That being said, when adding reserve actuators, they should act on the body as rarely as possible, to avoid getting incorrect simulation results.

My question is the following, in order to determine what optimal force to give to the reserve actuators, would it be a good strategy to give them the lowest possible value possible without the static optimization failing? That would mean running static optimization successively with "weaker" and "weaker" reserve actuators until the operation fails, and sticking with the smaller ones, or could there be a higher optimal force value that would yield more accurate results?

Also, the OpenSim documentation states:
Peak reserve actuator torques should typically be less than 10% of the peak joint torque. Peak residual forces should typically be less than 10-20 N; peak residual moments, less than 75 Nm (depending on the type of motion).
(https://simtk-confluence.stanford.edu:8 ... +residuals).

I hope my question is clear, thank you very much in advance!

Tags:

User avatar
Carmichael Ong
Posts: 387
Joined: Fri Feb 24, 2012 11:50 am

Re: Reserve actuators optimal force value

Post by Carmichael Ong » Sun Apr 19, 2020 10:07 pm

Yes, in general you want to have reserve actuators' optimal forces to be small. One other reason for this is how the static optimization problem is posed. If the optimal forces of the reserve actuators are high, then it's "cheap" to use in the cost function of static optimization. Thus, they should be small so that the muscles are used primarily to drive the motion, rather than the reserve actuators.

User avatar
Charles David Sasportes
Posts: 15
Joined: Thu Jan 23, 2020 6:36 am

Re: Reserve actuators optimal force value

Post by Charles David Sasportes » Sun Apr 19, 2020 11:30 pm

Thanks for your reply,

Do you know by any means where I can find info on the peak joint torque of my model so that I can make sure I satisfy the
Peak reserve actuator torques should typically be less than 10% of the peak joint torque.
condition ?

User avatar
Carmichael Ong
Posts: 387
Joined: Fri Feb 24, 2012 11:50 am

Re: Reserve actuators optimal force value

Post by Carmichael Ong » Mon Apr 20, 2020 1:25 pm

I believe that is referring to the peak joint torques for the motion (such as the output when you run inverse dynamics) rather than the peak joint torque of the model.

User avatar
Charles David Sasportes
Posts: 15
Joined: Thu Jan 23, 2020 6:36 am

Re: Reserve actuators optimal force value

Post by Charles David Sasportes » Wed Apr 22, 2020 6:24 am

many thanks !

User avatar
João Belo
Posts: 1
Joined: Sun Nov 24, 2019 6:55 am

Re: Reserve actuators optimal force value

Post by João Belo » Thu Apr 23, 2020 12:39 am

Hi, I am currently trying to use the MObL_ARMS model to solve some static optimization problems on different static arm poses. I currently have a reserve actuator for each coordinate in the model with a small optimal force of 1. However, for each timestamp in my motion, static optimization gives very different results.
From my understanding, the optimizer is converging to different solutions, and some of them have really high "loss/cost/performance".
Am I doing something wrong regarding the actuators? Is it possible to make it more consistent somehow?
Thanks for this discussion in advance!

User avatar
Carmichael Ong
Posts: 387
Joined: Fri Feb 24, 2012 11:50 am

Re: Reserve actuators optimal force value

Post by Carmichael Ong » Tue May 05, 2020 5:13 pm

Unsure of what might be going on there. Seems like the optimizer and/or simulation is caught in some weird state. Do you have a minimum reproducible example?

POST REPLY