[ODE] Tri-Tri from a slightly different angle...

Shaul Kedem shaul_kedem at yahoo.com
Wed May 5 10:22:45 MST 2004


Hi all,
 I just had an idea, which might be either genius or..
well.. not.. 

 The problem at hand is how to collide two tri-meshes.


My idea is to fill a tri mesh with balls prior to
running ODE, make a composite object out of them and
then run a two step process:
 1. collide the composite objects.
 2. use the tri mesh to calculate the actual angle of
collision..

 Here is some pseudo ascii art:

 +------+  +------+
 |B/--\O|  |B/--\O|
 | |Ba| |  | |Ba| |
 |X\--/ |  |X\--/ |
 +------+  +------+

 Where "BOX" is a 3 dimensional box and "Ba" are the
composite object, in our case a ball...

 As you can see(or can you.. heh), the ball touches
each of the faces in a certain point, which makes the
"in" and the "out" of the mesh very easy to figure
out... there is only one side in which the ball
touches the face.. that is "inside".
 
 To do that an algorithm should be created (invented?)
which "fills up" a tri-mesh with balls... basically
the algorithm should find the largest space it can
fill a ball in (touching the highest number of faces)
and go on until the whole trimesh is filled... also, a
"fudge" factor can be *safely* calculated, this fudge
factor will then be used to find a valid radius of
faces in case a certain ball doesn't actually touch a
face...

  So, my questions are:
  a) Will this method work?
  b) have anyone made such an algorithm yet?
  c) What have I forgot... ?

(If this isn't a new idea, my apologise..)

Best,
Shaul


	
		
__________________________________
Do you Yahoo!?
Win a $20,000 Career Makeover at Yahoo! HotJobs  
http://hotjobs.sweepstakes.yahoo.com/careermakeover 


More information about the ODE mailing list