[ODE] High speed rigid body collisions

[Darío Mariani]

  A quick thought, you can use a caped cylinder instead of a ray, in
this way you do not have the problem with edges (even you can have
holes on the wall). Your problem will be how to find the position
where the collision occurs, I think two approaches could be used, 1)
Do some math to find where they would collide, 2) Move the ball along
the path and try to use something to reduce the iterations to a
minimum.

         Darío


On 6/1/05, Zulfiqar Malik <Zulfiqar.Malik at itrango.com> wrote:
> I think that the ray casting approach would fail in a lot of cases
> especially when you have a scenario when the ball is slightly large
> (consider football or basketball) and the wall is finite such that the
> ball can collide with the edges of the wall. In that case your ray will
> miss the wall (most of the time, if not always) and your ball will again
> tunnel through the wall! I don't exactly whether such an effect would be
> feasible for you or not (e.g. it might be feasible for a wall knocking
> sort of tennis game), but in majority of game/simulation environments,
> this would be pretty common and would cause undesired results especially
> if you have a large sphere with finite flat objects all around.
> 
> What you can do is to write your own tri-mesh collision routine. It's
> pretty easy and there are plenty of tutorials online, especially for
> spherical/elliptical objects.
> 
> Regards,
> Zulfiqar Inayat Malik.
> 
> _______________________________________________
> ODE mailing list
> ODE at q12.org
> http://q12.org/mailman/listinfo/ode
>