Stickbot feet fixed enters hardware fault

[GrmanRodriguez]

After implementing this PR the stickbot_feet_fixed model enters hardware fault immediately:

The main reason seems to be these 4 links:

  <link name="base_link"/>
  <link name="l_sole"/>
  <link name="r_sole"/>
  <link name="imu_frame"/>

We would like to understand exactly what is happening and know if we can still get away with removing the inertial for these links

[GrmanRodriguez]

Upon further tests I've noticed the main culprit is base_link. Removing the inertial of this link causes the hardware fault.

The weird thing is base_link is connected only to root_link by a fixed joint, so it's not affecting other elements of the kinematic chain.

The joints in fault are r_hip_pitch, l_hip_pitch (both connecting root_link with the respective hip_1) and, strangely, torso_pitch (which connects torso_1 and torso_2).

root_link seems to be similar to its icub-models counterpart, so it's not clear yet why there is the fault.

cc @Nicogene