In the solution and urdf files for Mira’s head the inertia values used are based on a cylinder but geometry is defined as a sphere. I noticed that at one time a cylinder was used and is now commented out in is the inertia just a left over from when a cylinder was used. If not why use a cylinder for inertia and sphere for the geometry ? I know for a collision using a cylinder makes sense so you can avoid any contact with the head which contains the sensors. About to take quiz and I just want to know before taking it.
I presume you are talking about this code:
<link name="head_link">
<inertial>
<origin xyz="0 0 0" rpy="0 0 0"/>
<mass value="0.02" />
<inertia ixx="2.61666666667e-05" ixy="0.0" ixz="0.0" iyy="2.61666666667e-05" iyz="0.0" izz="3.6e-05"/>
</inertial>
<collision>
<origin xyz="0 0 0.015" rpy="0 0 0"/>
<geometry>
<sphere radius="0.06"/>
</geometry>
</collision>
<visual>
<origin rpy="0.0 0 0" xyz="0 0 0"/>
<geometry>
<sphere radius="0.06"/>
</geometry>
</visual>
</link>
Definitely that is wrong. If the geometry within the collision tag is a sphere, then the inertia must be computed for that (pay attention that the geometry tag within the visual part doesn’t matter at all since that is just for visualization purposes, nothing related with the physics of the robot)
As we can see in the code, the inertia tag is being computed for a cylinder:
I_xx = I_yy = (1/12) * m * (3 * r^2 + h^2)
I_zz = (1/2) * m * r^2
Instead, it should be computed for a sphere:
I_xx = I_yy = I_zz = (2/3) * m * r^2