Jacobsen & Verlet - embedding simplex in polytopes
Posted: Sat Jul 25, 2015 1:58 pm
Hi there,
Like loads of other people, I've been tinkering away with Jacobsen's verlet approach (http://web.archive.org/web/201001110352 ... dc2001.htm) for a little game for my kids.
I have working tetrahedrons bouncing around and interacting on a heightmap, so far so good. However, what confuses me is Jacobsen's description of embedding tetrahedrons in more complex rigid bodies.
As far I understood, the collision detector is supposed to detect collisions on the wrapping rigid body. OK, not a problem. Then, and here I'm lost, one is supposed to use the collision response (penetration distance and normal) on the embedded tetrahedron. What of the collision point? The point cannot be described by barycentric coordinates of the tetrahedron as it is on the outside, thus I cannot use the Lagrangian multiplier to update the tetrahedron particles. Am I supposed to project the collision point onto the tetrahedrons surface or something?
Anyone worked this out? I've been googling around like crazy last week but fail to find anything. I even bought the book Game Physics Pearls, but Jacobsen's article here was much the same as the link above (Otherwise, a very good book).
Thanks / Peter
Like loads of other people, I've been tinkering away with Jacobsen's verlet approach (http://web.archive.org/web/201001110352 ... dc2001.htm) for a little game for my kids.
I have working tetrahedrons bouncing around and interacting on a heightmap, so far so good. However, what confuses me is Jacobsen's description of embedding tetrahedrons in more complex rigid bodies.
As far I understood, the collision detector is supposed to detect collisions on the wrapping rigid body. OK, not a problem. Then, and here I'm lost, one is supposed to use the collision response (penetration distance and normal) on the embedded tetrahedron. What of the collision point? The point cannot be described by barycentric coordinates of the tetrahedron as it is on the outside, thus I cannot use the Lagrangian multiplier to update the tetrahedron particles. Am I supposed to project the collision point onto the tetrahedrons surface or something?
Anyone worked this out? I've been googling around like crazy last week but fail to find anything. I even bought the book Game Physics Pearls, but Jacobsen's article here was much the same as the link above (Otherwise, a very good book).
Thanks / Peter