Muscle activations/forces in CMC and Static Optimization
Posted: Fri Jun 27, 2014 5:46 pm
Hi,
I am interested in examining muscle forces and activations during human running. I have worked quite a bit with CMC but am now starting to experiment with possibly using static optimization instead in order to save computational time. My goal is to alter certain muscle properties and see the compensations that arise in the other muscles. I am also working on developing a new model with additional muscles so being able to compare muscle activations to EMG in order to validate the model is also important to me.
One thing I do not understand about Static Optimization is that when I run the Hamner Full Body Model (2010) through CMC and then the exact same model (and same RRA input) through SO, the magnitude of the muscle activations and in turn the muscle forces in SO are significantly smaller than in CMC, although the patterns are still somewhat similar. I am aware SO solves each frame independently in time but I dont understand how with the same motion, model, ground reaction forces, etc model activations and forces could be so different (at least 5x smaller muscle forces). Static optimization is minimizing the sum of the activations squared -- isnt this similar to what CMC is optimizing also? Does anyone know why this might be happening? Anderson et al 2001 showed that dynamic and static optimization solutions for gait produced very similar estimates of muscle forces, so I am wondering if I am possibly doing something wrong with SO? Any help or insight would be very appreciated!
Thank you!
I am interested in examining muscle forces and activations during human running. I have worked quite a bit with CMC but am now starting to experiment with possibly using static optimization instead in order to save computational time. My goal is to alter certain muscle properties and see the compensations that arise in the other muscles. I am also working on developing a new model with additional muscles so being able to compare muscle activations to EMG in order to validate the model is also important to me.
One thing I do not understand about Static Optimization is that when I run the Hamner Full Body Model (2010) through CMC and then the exact same model (and same RRA input) through SO, the magnitude of the muscle activations and in turn the muscle forces in SO are significantly smaller than in CMC, although the patterns are still somewhat similar. I am aware SO solves each frame independently in time but I dont understand how with the same motion, model, ground reaction forces, etc model activations and forces could be so different (at least 5x smaller muscle forces). Static optimization is minimizing the sum of the activations squared -- isnt this similar to what CMC is optimizing also? Does anyone know why this might be happening? Anderson et al 2001 showed that dynamic and static optimization solutions for gait produced very similar estimates of muscle forces, so I am wondering if I am possibly doing something wrong with SO? Any help or insight would be very appreciated!
Thank you!