Research On Collision Detection In Virtual Environment

From the 1990s, with the rapid development of computer networks, virtual reality technology has been rapidly developed and strongly supported. In the field of virtual reality, collision detection is an important aspect of the application. Accurate and fast collision detection can improve the reality and immersion of the virtual environment, but for the virtual environment itself, the complexity and real-time characteristic bring higher requirements to collision detection.With the background of the design to a virtual assembly system and aiming to different kinds of objects using different methods of collision detection, this paper designed two collision detection algorithms, one is based on bounding box hierarchy solving rigid body collision, the other is based on niche PSO solving deformable body collision.Bounding box hierarchy is an effective solution to the problem of rigid body collision detection. It describes approximately the object with a simpler bounding box that is a litter bigger than the object, and detects whether there has a collision between bounding boxes by the level of binary tree traversal, and then continues to do the intersection test to the basic geometry elements inside the box to determine whether there has a collision between objects. The particle swarm algorithm of Swarm Intelligence Algorithm was introduced to the field of collision detection, combined with the niche and sampling technology, broke through the limitations of topology, solved the problem of collision detection of the deformation body in the dynamic peak environment. In this paper, two algorithms have been programmed to verify the effectiveness of collision detection, the result of experiments showed that the both algorithms can detect collision accurately and fast.The paper use the both algorithms to detecte dollision in the virtual assembly system. The results further proved the correctness and accuracy of the algorithm.