home *** CD-ROM | disk | FTP | other *** search
- Dimitri Holman wrote:
-
- > Collision Detection in 3D is non-trivial BUT can be done fast with the
- > correct algorithm. Fast as in REAL-TIME. We use it in the Vis-Sim world
- > all the time.
-
- Is that why so many of the DIS runs I've seen feature "flying" and "bur-
- rowing" tanks? Real-time 3D surface interaction isn't very robust.
-
- Ted Stethem replied:
-
- > In 3D, it would be necessary to have X, Y and Z coordinates for the
- > points plus have the capability to interpolate points in-between when
- > intersecting between them, say a line between two sets of coordinate
- > points. I need to dig up my calculus book and go over the section
- > about intersecting surfaces again.
-
- It's worse than that. If your terrain is smoothed or bump mapped, for
- example, you have to calculate the interaction based on the continuous
- surface implied by the shading, and not the geometry of the underlying
- polygons. And consider a wheel rolling over a peak: the axis of rota-
- tion moves and the wheel rotates, but the point of contact with the
- terrain is fixed for a time.
-
- I doubt the calculus book will be much help, since most problems of this
- kind are difficult to solve analytically. Look for books on numerical
- methods in physical simulation, in addition to material specific to CG.
-
- - Ernie
-
-