[ODE] Dynamic mesh collision

Amund Børsand amund at c2i.net
Tue Feb 18 12:28:02 2003


(I'm sorry I privatesent this message to you, it was meant to go to the
list. I also edited it slightly after figuring out who you were ;)

"Pierre Terdiman" <p.terdiman@wanadoo.fr> skreiv:

> > Yo, I suddenly thought of something. I have an idea about implementing
> > "real" tyres in my offroad simulator, as a cloth-like mesh with inside
> > pressure. This mesh would then have to collide with the ground. But I
> > remember reading somewhere that OPCODE don't work with
> > deformable/dynamic meshes, is that right? And the tricollider use OPCODE,
> > right?
> >
> > Hm. Problem, then.

> If the tyre is a mesh and the ground is a mesh, you have more serious
> problem than the deformable nature of the tyre (Opcode 1.3 (unreleased)
> supports dynamic meshes, for example). Here you would need mesh-vs-mesh
> collision, which is difficult enough just for the static case....

First of all I'm impressed with OPCODE, and looking forward to the 1.3
relase.

What I thought is that since the tyre mesh probably will have fairly
high resolution compared to the terrain, and the rubber is supposed to
be flexible between the points anyway, so point-against-mesh checking
would be nice, I think. I'd need to calculate the forces back to each
vertex anyway, to deform the mesh and translate forces back to the rim.

Is it difficult to do point-vs-mesh checking? Wouldn't that just be a
simplification of sphere-vs-mesh? But problems may arise when I have to
check a lot of individual points against the same mesh, maybe?

> I'd forget it.

I can't.