Dynamic Collision Detection in Virtual Reality Applications

We present data structures and algorithms for dynamic collision detection in virtual reality (VR) applications. The methods are applicable to all general polygonal models. They combine the advantages of collision detection using bounding volume (BV) hierarchies with the ability to compute dynamic collision detection results. The results are used as input for further simulations, e.g. contact or dynamics simulation. First we present new methods to compute BV hierarchies using optimization goals which can also be used to improve known computation methods. Second we show how to integrate BV hierarchies into a process of dynamic collision detection, so that the bounding objects as well as the surface patches of the objects are tested for overlap during their motions. The performance of the techniques is shown by means of a fitting simulation in the automotive industry.