Screening of factors influencing the performance of manipulator using combined array design of experiment approach

Abstract

A robot must manipulate objects with high accuracy and repeatability to perform precise tasks. However, deviation in performance is attributed to uncertainties and improper selection of control, noise, and process factors. The information regarding the effect of these factors on performance is almost non-existent. A probabilistic approach has been used to model and simulate the performance of manipulator. The combined array fractional factorial design of experiment approach has been employed to identify the significant factors and their interactions. This approach helps in screening of the manipulator factors and focus on those that are important. To explore further, two indices, viz. link length ratio and link mass ratio, have been proposed and impact of these indices on manipulator performance is investigated. A two degree of freedom (2-DOF) RR planar manipulator performing a task with cubic and quintic trajectory has been used to illustrate the approach. It has been observed that the statistically significant factors are different for different tasks in workspace. It has also been observed that for the same task, factors responsible for performance variations are different for cubic and quintic trajectories. Finally, it has been demonstrated that the link length ratio change has significant influence on performance compared to link mass ratio.