study the impact of exoskeleton on energy consumption Opensim

[Nafiseh Ebrahimi]

Hello everyone,
I am new in Opensim and my project is about study and simulation of the impact of exoskeleton (external actuators) on energy consumption. I want to add actuators on hip, knee and ankle and see how they affect walking, running and required energy? is this simulation doable in opensim? how?
I would appreciate if somebody can help me in this regard.

Thanks
Nafis

[Dimitar Stanev]

Hi,

Yes it is doable. A good place to start is:

http://simtk-confluence.stanford.edu:80 ... %27s+Guide

For energy consumption you can see the following tutorial:

http://simtk-confluence.stanford.edu:80 ... bolic+Cost

There is a lot of material and tutorials on the user guide site, so spend some time to study it.

Good luck

[Christopher Dembia]

You might be interested in this recent paper led by Tom Uchida:

http://journals.plos.org/plosone/articl ... ne.0163417

[Nafiseh Ebrahimi]

Thank you for your Help. I have one more question. in the link below:
http://simtk-confluence.stanford.edu:80 ... bolic+Cost
when I run the CMC tool I can see the model walking which is muscle driven. So I am thinking that the muscle activation makes the model to run. But when I disable whole muscles of one leg the model still walks without any limping and problem. I cannot understand what happens. Do you know what is the reason?

Hi,

Yes it is doable. A good place to start is:

http://simtk-confluence.stanford.edu:80 ... %27s+Guide

For energy consumption you can see the following tutorial:

http://simtk-confluence.stanford.edu:80 ... bolic+Cost

There is a lot of material and tutorials on the user guide site, so spend some time to study it.

Good luck

[Thomas Uchida]

CMC drives the model through an experimentally observed motion and computes the muscle forces that would have generated that motion; disabling a muscle will not change the motion it is tracking.

The same question was asked on another thread (viewtopicPhpbb.php?f=91&t=3675&p=19534); please see my responses there and the documentation for CMC here: http://simtk-confluence.stanford.edu:80 ... le+Control.

[Thomas Uchida]

The so-called "reserve actuators" may be a source of confusion. Here is an excerpt from the "How CMC Works" page on Confluence (http://simtk-confluence.stanford.edu:80 ... +CMC+Works):

"The fast target is both faster and generally produces better tracking. However, if the constraints cannot be met, the fast target will fail and CMC will exit with an error message. Often the reason for the failure is that the musculoskeletal model is not strong enough.

To prevent the fast target from failing, it is possible to add a number of reserve actuators to a model that are able to make up for strength deficiencies in muscles should any be encountered. The reserve actuators have very low strength (or optimal force) and so require very high excitations to apply substantial load to the model. As a result, use of the reserve actuators is highly penalized in both the fast and slow formulations. When the forces (or moments) applied by the reserve actuators are written to file and plotted, their values are a good indication of which joints in the musculoskeletal model are not strong enough. If the model is strong enough, the reserve actuator forces/moments should generally be very small relative to the forces/moments applied by the primary actuators."

[Nafiseh Ebrahimi]

CMC drives the model through an experimentally observed motion and computes the muscle forces that would have generated that motion; disabling a muscle will not change the motion it is tracking.

The same question was asked on another thread (viewtopicPhpbb.php?f=91&t=3675&p=19534); please see my responses there and the documentation for CMC here: http://simtk-confluence.stanford.edu:80 ... le+Control.

So, how Opensim can help us to simulate movement of people with some disabilities and weak muscles. I think after doing CMC and obtaining muscle excitation we must use excitation Editor and change a specific excitation which is a number between 0-1. after that, use this edited file to run Forward Dynamics and this tool can show us created faults in walking when we play model motion in GUI. Am I right?

[Nafiseh Ebrahimi]

in the link below related to metabolic costs:
http://simtk-confluence.stanford.edu:80 ... bolic+Cost

it is mentioned that : "the OpenSim software platform has an extensive library of actuators and controllers."

How we can use this library in order to add External actuators (assistant torques) to the model's joints. Is it possible without using XML Marker or Scripting? is there any helpful document in this regard?

Thanks

[Thomas Uchida]

So, how Opensim can help us to simulate movement of people with some disabilities and weak muscles. I think after doing CMC and obtaining muscle excitation we must use excitation Editor and change a specific excitation which is a number between 0-1. after that, use this edited file to run Forward Dynamics and this tool can show us created faults in walking when we play model motion in GUI. Am I right?

That strategy is unlikely to generate satisfactory results. First, you would also need to define a contact model between the feet and ground to generate GRFs in your forward simulation. Second, unless you've also implemented some sort of balance controller, the model will fall because it's an unstable system; the "Dynamic Walking Challenge" example (http://simtk-confluence.stanford.edu:80 ... Id=5113821) illustrates this well.

It isn't clear what you mean by "how Opensim can help us to simulate movement of people with some disabilities and weak muscles". If you've collected data from subjects with muscle weakness, you can run through the typical pipeline using an appropriately adjusted model (e.g., with lower values for the peak isometric muscle forces). If, on the other hand, you want to predict the kinematics of a person with muscle weakness (i.e., in the absence of experimental data), then the situation is more complicated. Here's one approach:

Dorn, T.W., Wang, J.M., Hicks, J.L., Delp, S.L. Predictive simulation generates human adaptations during loaded and inclined walking. PLOS ONE 10(4), 2015. http://nmbl.stanford.edu/publications/pdf/Dorn2015.pdf

it is mentioned that : "the OpenSim software platform has an extensive library of actuators and controllers." How we can use this library in order to add External actuators (assistant torques) to the model's joints. Is it possible without using XML Marker or Scripting? is there any helpful document in this regard?

There are at least three ways of adding components to an OpenSim model:
- Modifying the XML directly. In the GUI, select "XML Browser" from the "Help" menu to get a template for the XML code. The API documentation provides descriptions of the actuators (e.g., CoordinateActuator, https://simtk.org/api_docs/opensim/api_ ... uator.html) and controllers (e.g., PrescribedController, https://simtk.org/api_docs/opensim/api_ ... oller.html) that are available.
- MATLAB/Python scripting interface. See the "Scripting" section in the Confluence documentation (http://simtk-confluence.stanford.edu:80 ... /Scripting) for details and example scripts.
- Using the C++ API. There are examples here: http://simtk-confluence.stanford.edu:80 ... I+Examples.

[Nafiseh Ebrahimi]

Hi,

Yes it is doable. A good place to start is:

http://simtk-confluence.stanford.edu:80 ... %27s+Guide

For energy consumption you can see the following tutorial:

http://simtk-confluence.stanford.edu:80 ... bolic+Cost

There is a lot of material and tutorials on the user guide site, so spend some time to study it.

Good luck

In this tutorial when I plot for muscle forces in assisted gait model I can see a muscle named, PlantarFlextionAssist. I assume that this extra muscle which is absent in unassisted model is the actuator which is applied to the generic model. Is this assumption true?