Simultaneous optimization of timing and trajectory in sequential and parallel tasks of humanoid robots

Abstract

This paper presents a motion planning technique capable of scheduling multiple sub-tasks that can be achieved in different orders and at various timings. A robotic motion is defined as a timed event sequence of multi-body dynamical system. Schedules of the motion planning is determined by robotic mechanism and motion requirements. Motion planning problem including task scheduling is then formulated as a class of constrained optimization problems, which can be solved efficiently by means of an iterative algorithm that exploits the sparsity of the multi-body system. The proposed method is demonstrated in object-catching tasks of a standing humanoid in numerical simulation.