确定智能机器人抓手姿态的几何推理算法

The Geometric Reasoning Algorithm for Grasping Posture of Intelligent Robot

介绍了一种求解智能机器人抓取规划的几何推理算法。该算法基于平面多面体表示的物体几何模型,对物体上的各类几何项(即面、边及顶点)之间的空间关系进行分析,求解出所有无碰撞可抓取的抓手姿态参数(即位置矢量、接近矢量、方位矢量及开度)。并以抓取类型的稳定性和物体重心的空间位置作为启发信息,搜索出机器人抓手在物体上的最优抓取位置,进而确定机器人抓手的最优姿态。

In this paper, a geometric reasoning algorithm for grasping planning of intelligent robot is introduced.Based on the geometric model of plane polyhedron of object,this algorithm can search out all possible collision-free grasping parameters(i.e vectors of position, approach and orientation and opening distance) by analysing the space relationship among the geometric items (i.e face, edge and vertex). According to the heuristic information of grasping stability and the center-of-gravity position of object,it can get the best grasping position and posture of manipulator.