Some desirable capabilities

Provide easy-to-use, extensible software for modeling, simulating, controlling, and analyzing the neuromusculoskeletal system.
POST REPLY
User avatar
Marko Ackermann
Posts: 1
Joined: Fri Jun 08, 2007 1:35 pm

Some desirable capabilities

Post by Marko Ackermann » Mon Nov 12, 2007 8:50 am

Dear OpenSim developers,

I am a member of the project “Multidomain Efficient Methods for Multi-Domain Biomechanical Simulations” leaded by Ton van den Bogert and Ahmet Erdemir. Our group is very interested in actively using OpenSim in our musculoskeletal system simulations and optimizations. However, we miss some capabilities, which would greatly increase the usefulness of OpenSim for our project. Below I list some of the desirable capabilities, many of them are probably already implemented, and just in need of documentation.

1-) A programming interface (API) which gives us access to the musculoskeletal model. This should be in the form of a set of differential equations: dx/dt = f(x,u). There should be a function we can call with inputs x and u, and the function should compute dx/dt. x are all state variables, kinematic states and muscle states. u are the control inputs, i.e. the neural excitations.

2-) A plugin interface which gives us a GUI to run our code. Our code would use the OpenSim API internally, for accessing the model and for visualization and plotting.

3-) An interface which allows us to use our own ground contact model. OpenSim would call our function, provide bone position and orientation, and we would return reaction loads.

4-) An interface which allows us to use our own muscle path model. OpenSim would call our function, provide skeleton generalized coordinates, and we would return muscle length and moment arms. We could call OpenSim's internal function for doing this job, but this may be slow. We consider having our surrogate modeling engine in this interface, which saves these results in a database and then gradually manages to avoid doing the actual muscle path calculations.

5-) We could use OpenSim’s internal Hill muscle model, but it would be useful to be able to plug in our own muscle model, e.g. to avoid discontinuities.

Thank you,

Marko Ackermann


POST REPLY