Monday, August 07, 2006

Picking up the flocking algo once again.


I went to the SIGGRAPH convention in Boston. There I ran into probably the only person who remembers my work with flocking algorithms in povray. With this encouragement, I started dusting off the code.

Here is a recent test of the combination of three goals: i) following after a moving goal particle; ii) avoiding collisions with neighbors, and iii) avoiding collisions with the objects (here, two planes and some rows of cylinders (which mysteriously don't show up in this orthographic view). The image shows the positions of about fifty particles after some 200 frames. Some (in blue) have nicely negotiated the maze of cylinders I've set before them.

The problem i'm having is that a lot of the particles are running out through the floor of the set. I'm not sure what's causing it, as it appears to get worse as I increase the repulsion with objects in the scene. More to follow.

No comments: