TR-IIS-03-005    PDF format

Abstract

This paper is aimed at introducing a faster collision detection algorithm for convex polyhedral in a three-dimensional workspace. The problem of collision detection has been researched in the literature of computer graphics, robotics, computational geometry, computer animation, and physically-based modeling. It has been regarded as a computationally demanding task and is often treated as an advanced feature. In this paper, we present a faster way to solve the collision detection problem by using the enclosed ellipsoid method. Enclosed ellipsoid method gets fewer collision error report than the standard bounding volume method, such as boxes or spheres.  We split ellipsoids into facial areas and ignore the areas which are always the heavier computational load but impossible to be collided with. There are some computer simulations showing how the facial split method increases the efficiency of the detection in comparison with the original enclosed ellipsoid method.