RE: [Algorithms] Colliding against polygon soup

[Andrew F. <An...@cl...>]

if you can afford the memory i think it's quite common to use obb trees

[see Gottschalk, Lin and Manocha's paper: OBBTrees: A Hierarchical Structure
for Rapid Interference Detection].

also you might wanna look at Pierre's optimised form which reduces memory
usage to about 80 bytes per tri [if your tree goes down to individual tri
level]

http://www.codercorner.com/RAPID_Hack.htm

i'm personally working with poly soup in a low[ish] mem environment on an
rpg [read: unrealistic physics] stylee game. Walking about on a landscape,
throwing objects, being hit by arrows [or railguns:)], driving around, etc.
Never likely to be particularly many units around at any one time. i get by
with good spatial hierarchy, sensible cacheing of previous data, and use of
bv's to trivially reject as much as pos. Then if a potential collision _is_
detected just do a [low lod obviously] poly-poly check. it's a suitable
solution for the type of game.

btw i think there's a fair amount on collisions in the archives [if they're
working...]

HTH

Andy

8= -----Original Message-----
8= From: Brian Hook [mailto:bw...@wk...]
8= Sent: Tuesday, December 05, 2000 06:46
8= To: gda...@li...
8= Subject: [Algorithms] Colliding against polygon soup
8= 
8= 
8= Curious what strategies people are using to collide against polygon 
8= soup.  I know a lot of people are building levels these days 
8= using general 
8= polygon modelers such as Maya and MAX, and these products 
8= don't really 
8= support a notion of "in" or "out", so you have to simply watch for 
8= transitions into a polygon.
8= 
8= I don't have ANY experience with collision detection, so I'm 
8= curious what 
8= direction I should go (no pun intended).
8= 
8= Brian
8= 
8= _______________________________________________
8= GDAlgorithms-list mailing list
8= GDA...@li...
8= http://lists.sourceforge.net/mailman/listinfo/gdalgorithms-list
8=