Global optimization of motion planning for a flexible macro-micro manipulator system using Estimation of Distribution Algorithm

* Supported by the National Natural Science Foundation of China(No. 60305008 ) Abstract This paper proposes a new method for motion planning global optimization of a macro-micro manipulator (3M) system based on Estimation of Distribution Algorithm (EDA). Due to the redundant character of a 3M system, its motion planning is formulated as a multi-objective optimization problem under the mechanical constraints. After a task space division, optimization parameters are transferred from joint angles to division points of macro and micro part of the system, which largely reduces the searching space. However, to avoid sucking local extremum, global optimization is too time-consuming to bear. Instead of searching the division point space directly, EDA is applied to estimate the probability model of the division space and thus extracts mapping information between the optimization function and division points. Simulation results on a planar 4 degree-of-freedom (DOF) macro-micro manipulator system are presented herein to show the effectiveness of the proposed methods.