Comparison cuboid muscle vs biceps muscle

[Lukuas31415]

We have been using cuboid muscles for most examples to see how dummy muscles behave. We have also created a biceps case for a more realistic model. Unfortunately, these two models create very different results. The following cuboid muscle kind of resembles the biceps:

This is not perfect, but close enough that the muscles can be compared when being activated. If we use Bayesian Optimization on those muscles, we get the following plots:

Biceps:

Cuboid:

As we can see, the muscles behave extremely differently. Not only does the biceps have a non-trivial maximum and the cuboid doesn't, but also the values of the cuboid are about 15 times higher than the biceps'. If we take the norm of the traction vectors instead of just the z-component, the biceps has slightly higher values and is also strictly monotonically increasing, but the cuboid values are still by magnitudes higher. Question is, why this happens.

[Lukuas31415]

The exact setups of the biceps and cuboid muscles can be found in the branch comparison_biceps_cuboid.