Manipulation Planning Algorithm Using Preferential Attachment of Subgoals Based on Dynamic Multi-fingered Manipulability
Masahito Yashima · Transactions of the Society of Instrument and Control Engineers · 2006
This paper presents a manipulation planning using bias sampling of subgoals based on multi-fingered manipulability. The randomized planning we have proposed is based on random generation of subgoals for object configuration, and the connection between these subgoals without considering the manipulability. Therefore, the randomized planning increases calculation time and trial frequency. In order to improve these problems, in the present algorithm, we first choose a high manipulability subgoal and generate a next subgoal in the area biased for target direction. We verify the effectiveness of the present algorithm by simulation examples.