Study On Collision Detection Algorithm Of Hybrid Bounding Box

Collision detection is pretty important in Virtual Reality, and its speed greatlyinfluences the reality and illusion of immersion in virtual environment. The algorithms ofcollision detection are studied in this paper deeply, which contains the following parts:Firstly, the present situation and the techniques of virtual reality and collisiondetection are reviewed briefly. Then several common bounding volume hierarchy (BVH)algorithms including AABB, OBB and k-DOPs are investigated systematically. Next,these BVH methods are compared and analyzed.AABB BVH is easily updated and adapts to deformable objects. The updating speedof OBB BVH is slow, but its bounding box is tight and suited for solid objects. In view ofthese virtues, the writer produces his own collision detection algorithm—hybrid boundingbox method, which applies OBBS to solid objects and uses AABBs to deformable ones.This method can be well used for those complicated environments filled with solid anddeformable objects. Next the existing problems in this hybrid method are explained anddemonstrated in detail, which include the construction of BVH, the update and theself-intersection detection of BVH. And the writer produces an intersection detectionalgorithm between AABB and OBB bounding boxes according to the separating-axistheorem. And the hybrid algorithm can adapt to the changes of the topological structureafter objects are deformed.In the following part, the hybrid bounding box algorithm is applied to the experimentof surgery simulation. And this method is also applicable to different surgical tools. It isan efficient algorithm to solve collision detection problem in complicated scene. Severalspecial intersection detections are also implemented including between line and cuboidand between sphere and cuboid. With certain experiments it is proved that this methodcan promote the efficiency of collision detection and realize real-time effect during theinteractive organ incision.