Implementation of vision-based planar grasp planning

摘要

This research describes the implementation of a vision-basednalgorithm that is capable of rapidly determining robotic grasp pointsnfor planar objects. A representation of the target and a quadtreenexpansion generate candidate grasps that are compared using a costnfunction. The approach returns the first acceptable grasp point at angiven tree resolution. The system has an execution time on the order ofnseconds and it is suitable for a large number of planar or near planarnobjects