Feasibility of disassembly tasks considering a posture of a subassembly using genetic algorithm

Abstract

Summary form only given. An automatic generation of assembly or disassembly sequences is a significant problem for assemblies using robot manipulators. We need to judge the feasibility of assembly/disassembly tasks based on geometric information of products. In previous research, only translational motion of a subassembly has been considered. However, there are many cases in which a subassembly can be separated from another subassembly using rotational motion. In this paper, we propose a new algorithm with which we can search a posture of a subassembly to avoid collision with another subassembly. The algorithm is based on the genetic algorithm. The proposed method makes it possible to search a posture of a subassembly to realize assembly/disassembly tasks with small computational amount.