自适应细分及优化编码八叉树碰撞检测算法

基于自适应三角网格划分和优化编码八叉树结构,在机器人虚拟手术训练系统中提出一种新的碰撞检测方法.采用该方法实现的八叉树结构占用存储空间小,且在具有良好的拓扑结构下又能保证实时性要求.八叉树结构中采用物体三角面片AABB包围盒平均边长的3倍设置八叉树空间单元大小,与已有文献采用包围球半径来设置八叉树空间单元大小相比,设置的单元大小合理有效,速度更快.最后根据三角形与最优单元大小相关性,通过自适应三角网格细分,利用三角形中心距离方法检测单元空间中三角形碰撞.实验仿真数据验证了该方法的实时性和有效性.

A Novel Algorithm for Collision Detection Based on Octree of Adaptive Subdivision and Encoding

Abstract： A novel algorithm for collision detection in robot operation training system based on octree of adaptive subdivision and encoding algorithm was proposed in this paper. Using this method, this structure can not only guarantee the real-time, and has the good topology. This octree structure adopts three times of the average triangle of the axis aligned bounding box to set the cell size. It is reasonable and effective compared with the radius of sphere. Based on the triangle size and the optimal cell size, by using the adaptive triangular mesh subdivision and the centroid distance, the collision in the space was detected. The experimental simulation data verify the real time and effectiveness of the method.