基于轴向包围盒碰撞检测算法的改进  

Mended Collision Detection Algorithm Based on Axis-aligned Bounding Box

作　　者：唐勇[1] 冯立颖[1] 吕梦雅[1] 

机构地区：[1]燕山大学信息科学与工程学院,秦皇岛066004

出　　处：《系统仿真学报》2009年第1期157-160,共4页Journal of System Simulation

基　　金：国家自然科学基金(50675189);河北省自然科学基金(F2006000267)

摘　　要：在虚拟环境中,活动对象的运动路径是连续的,为达到实时交互的目的,碰撞时间采样点的取值应该是十分密集的,因此,帧与帧之间会有很强的关联性。利用这一特性,对基于轴向包围盒碰撞检测算法进行优化,把当前对象的碰撞信息记录下来以供下一次碰撞检测使用。另外,针对AABB包围盒紧密性差、占用大量存储空间的问题,基于压缩存储的方法对该算法进行改进,减少AABB包围盒存储的字节数。实验结果证明,优化算法在有效地减少参与测试的包围盒数目的同时,也节省了AABB包围盒树的存储空间。In virtual environment, the paths of moving objects are continuous. In order to get the goal of real-time interaction, sampling points of collision time should be very dense. So there is stronger relevancy between frames. According to this trait, collision detection algorithm based on axis-aligned bounding box was optimized. The current collision information was recorded for using of the next collision detection. In addition, in order to solve the problem of bad tightness and storage space of AABB, the AABB algorithm was mended based on compressed storage method, and the optimizing algorithm reduced the amount of byte of AABB. The experimental results demonstrate that the algorithm has efficiently decreased the number of bounding boxes which participate in collision detection, and saved the storage space of AABB trees.

关 键 词：碰撞检测 轴向包围盒 时空相关性 压缩算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]