Planning of graspless manipulation based on rapidly-exploring random trees

Abstract

Planning of graspless manipulation is a difficult problem because of the complicated mechanics involved. The authors previously developed a motion planning method for general graspless manipulation by robot fingertips, but it requires a great deal of computation. In this paper, we present an improved method based on a probabilistic technique, RRTs (rapidly-exploring random trees), to accelerate planning. The method is applied to the planning of graspless manipulation where the manipulated object has three degrees of freedom, while our previous planner can deal with manipulation where the object has only one degree of freedom