You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I think the inertia ixx, ixy and izz was so much bigger than it should, the actual inertia should be about 0.000208 if mass was distributed equally within the cube , the original value was about 200 times bigger than it should, which mean the mass was distributed outside of the collision edge. Should change the inertia to <inertia ixx="0.000208" ixy="0.0" ixz="0.0" iyy="0.000208" iyz="0.0" izz="0.000208"/>
Hope someone better at math could confirm.
This is the cube.urdf file which was used in ShadowHand and AllegroHand training process.
I think the inertia ixx, ixy and izz was so much bigger than it should, the actual inertia should be about 0.000208 if mass was distributed equally within the cube , the original value was about 200 times bigger than it should, which mean the mass was distributed outside of the collision edge. Should change the inertia to
<inertia ixx="0.000208" ixy="0.0" ixz="0.0" iyy="0.000208" iyz="0.0" izz="0.000208"/>Hope someone better at math could confirm.