2023 年 41 巻 9 号 p. 801-804
In this work we developed a planner for planar sensorless in-hand caging manipulation for part feeders. In-hand caging manipulation is a method in which an object is caged by a hand throughout manipulation and located around a goal as a result of the deformation of the cage without sensing the object configuration. This method can be applied to part feeders to manipulate objects from random to targeted poses. In this paper, the following three points were proposed for part feeders, a new robot hand configuration named ``opposite-type hand'', improved motion planner for the hand and a hand closing algorithm. The first aims to make the manipulation more robust. The second is for computationally efficient planning of in-hand caging manipulation. The last improves pose accuracy by closing the robot hand until the object reaches desired pose. Experimental results demonstrated the effectiveness of the proposed points.