基于网格包络的工业机器人仿真碰撞检测算法

A Novel Collision Detection Algorithm Based on Grids Enveloping for Industrial Robot Simulations

为提高工业机器人在复杂作业环境下的碰撞检测效率,提出了一种网格包络的碰撞检测算法,以大量等尺寸的立方体网格来包络模型本身,并在网格内部建立网格子模型的AABB树结构。该算法在建模过程中将网格的空间坐标进行有序存储,在遍历阶段可快速搜索到相交的网格,之后遍历网格内部的树结构来进一步判断模型是否碰撞。该算法网格内部的子模型几何数据量远小于整体模型几何数据量,其网格内的检测速度远快于以整体模型建模的传统层次包围盒方法的检测速度。实验结果表明,在大型复杂模型碰撞检测仿真中,该算法在不同网格数量下的检测效率比传统的Solid算法的检测效率快数倍到数十倍。

To speed up the collision detection efficiency of industrial robots in the complex working environments,a novel collision detection algorithm was proposed using equal-sized cubic grids to cover the model and building tree structure of AABB in the grid.The space coordinates of these grids were stored orderly in the modeling progresses so as to determine whether there were grids intersecting in the traversal periods.Then traverse the hierarchical structure in the intersecting grids to detect collision precisely.Due to the grids had far less model data than the whole model,the detection speeds in the grids were far more fast than the traditional hierarchical bounding volume method where the building model was based on the whole model.The experimental results show that the detection efficiency in the novel algorithm is several times to dozes times（which depends on the size of grids）more than that in the traditional SOLID method for the large complex model environments.