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
There are computational advantages to collision detection using convex meshes. It would be useful for users who carefully construct their own convex meshes to be able to declare that they are convex. Then the collision engine can make smarter choices.
See also issue about gazebo-bullet concave mesh support for commentary on the challenges of real-time decomposition of concave meshes to convex. Using a flag will be computationally faster.
The text was updated successfully, but these errors were encountered:
If we don't want to impact runtime performance, we could add the flag and put the responsibility on users to use it correctly. We could supply an offline verification program to check that meshes declared as convex actually are
convex_hull - decomposes the mesh to one or multiple simple convex meshes. This can be done by the physics engine if available (e.g. bullet's btConvexHullShape) or use other libraries for decomposition, e.g. https://github.com/kmammou/v-hacd
bounding_box - Most simple option - just use the object's bounding box as the collision.
simple_shapes - decomposes to a set of simple shapes (instead of simple convex meshes)
Original report (archived issue) by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters).
There are computational advantages to collision detection using convex meshes. It would be useful for users who carefully construct their own convex meshes to be able to declare that they are convex. Then the collision engine can make smarter choices.
@erwincoumans @hsu
See also issue about gazebo-bullet concave mesh support for commentary on the challenges of real-time decomposition of concave meshes to convex. Using a flag will be computationally faster.
The text was updated successfully, but these errors were encountered: