Simultaneous selection of optimal parameters and tolerance of manipulator using evolutionary optimization technique

Abstract

For precise industrial applications, a manipulator must have high positioning accuracy and repeatability. However, variations in performance of the manipulator are attributed to improper selection of dimensions of design and process parameters and its tolerance. There have been almost no attempts to optimize these parameters and tolerance of the manipulator, by which performance variations will be minimum. This paper presents an offline approach to select optimal parameters and tolerance simultaneously which minimizes the manufacturing cost and delivers the target performance level. To determine the optimal parameters and tolerances, a modification in differential evolution algorithm is proposed, which helps in incorporating the effect of uncertainty in optimization process. Proposed approach has been illustrated by selecting optimal parameters and tolerance of a 2-DOF RR planar manipulator. The performance of this approach has been compared with modified genetic algorithm. It is observed that differential evolution algorithm provides optimal results with lesser computations compared to genetic algorithm.