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 spent a few hours debugging poor performance of FCL on some mesh models and realized that the broad-phase check wasn't pruning any pairs of geometries. I eventually tracked the problem down to incorrect AABBs on my CollisionObjects. It turns out that I wasn't calling computeAABB after changing the pose of my geometries, so the broad-phase check considered them all to be overlapping at the origin.
It would be helpful if you could add a note about this in the README. The current documentation doesn't mention computeAABB at all and, thus, gives the impression that the CollisionManager takes care of computing everything necessary for the broad-phase check.
(The performance is great now that the broad-phase check works!)
The text was updated successfully, but these errors were encountered:
I spent a few hours debugging poor performance of FCL on some mesh models and realized that the broad-phase check wasn't pruning any pairs of geometries. I eventually tracked the problem down to incorrect AABBs on my
CollisionObject
s. It turns out that I wasn't callingcomputeAABB
after changing the pose of my geometries, so the broad-phase check considered them all to be overlapping at the origin.It would be helpful if you could add a note about this in the README. The current documentation doesn't mention
computeAABB
at all and, thus, gives the impression that theCollisionManager
takes care of computing everything necessary for the broad-phase check.(The performance is great now that the broad-phase check works!)
The text was updated successfully, but these errors were encountered: