Hello all.
I have been trying to find studies using OpenSim in which the authors use open loop forward dynamics to validate control values found through CMC. I have thus far been unsuccessful. The papers I have found so far do use forward control, but it's part of the process in finding CMC values and then those values are assumed to be correct without running them through an open loop forward dynamics simulation to make sure the kinematics match up with those found using the IK tool.
Thanks for any help that anyone could give.
Brandon Brown
Studies involving open loop forward dynamics.
-
Brandon Brown
- Posts: 30
- Joined: Thu Jun 28, 2012 3:32 pm
-
Ton van den Bogert
- Posts: 164
- Joined: Thu Apr 27, 2006 11:37 am
Re: Studies involving open loop forward dynamics.
Brandon,
There is a recent paper that did exactly this test for an arm model:
Saul, KR et al. (2014) Benchmarking of dynamic simulation predictions in two software platforms using an upper limb musculoskeletal model. Comput Methods Biomech Biomed Engin [Epub ahead of print]
There is a PDF file on Kate Saul's website.
The forward simulation produced a movement that was very similar to the original movement that was used as a starting point, but the two drifted apart somewhat. If we assume that the exact same model was used for CMC and forward simulation, this has to be the result of the numerical approximations used in CMC and/or forward simulation. Some models and movements are more sensitive to numerical errors than others.
There are other ways than CMC to solve the inverse dynamic problem. One older (and slower) method is to repeatedly run forward simulations, with different controls, until you find the one that best matches the kinematics of the subject. This is known as a tracking optimization. Then, obviously, if you run a forward simulation with those controls, there will be no surprises. Even if the model is unstable and sensitive to the controls.
Ton van den Bogert
There is a recent paper that did exactly this test for an arm model:
Saul, KR et al. (2014) Benchmarking of dynamic simulation predictions in two software platforms using an upper limb musculoskeletal model. Comput Methods Biomech Biomed Engin [Epub ahead of print]
There is a PDF file on Kate Saul's website.
The forward simulation produced a movement that was very similar to the original movement that was used as a starting point, but the two drifted apart somewhat. If we assume that the exact same model was used for CMC and forward simulation, this has to be the result of the numerical approximations used in CMC and/or forward simulation. Some models and movements are more sensitive to numerical errors than others.
There are other ways than CMC to solve the inverse dynamic problem. One older (and slower) method is to repeatedly run forward simulations, with different controls, until you find the one that best matches the kinematics of the subject. This is known as a tracking optimization. Then, obviously, if you run a forward simulation with those controls, there will be no surprises. Even if the model is unstable and sensitive to the controls.
Ton van den Bogert