Search

A methodology based on Taguchi methods was implemented to determined/optimize robot process capability (RPC) for path following. The methodology consists of the characterization of the robot path data, the experimental design, the data analysis procedure, and the verification of the results. Four figures of merit (FOMs) for accuracy and four FOMs for repeatability were used to characterize a robot's controlled path capability. The experiment considered seven controllable factors (i.e. load, speed, shape, size, direction, orientation, and height) and used an L 27 (3 13 ) orthogonal array. From the results of the data analysis, the optimal factor-level combinations for the best path capability and minimum process variation were determined. The experimental runs were verified. The recommended methodology has the advantages of cost effectiveness, efficiency, data compression, and flexibility.

The robot process capability problem is outlined and terms are defined. Recent efforts to measure robot capability are reviewed. Two methods were used to analyze robot process capability data: the randomized complete block design method and Taguchi statistical experimental design methods. The purpose of the study was to compare the results obtained by these two methods and determine which method was better for analyzing the data to be used in determining/optimizing a robot's process capability. Three-dimensional (3-D) coordinate data were obtained using a WATSMART vision system to determine a robot's process capability. Three levels of four independent variables—load, X -, Y - and Z - coordinates—were chosen for the study. Both analysis methods yielded similar results in determining a robot's process capability (within 10.4%). The Taguchi methods required significantly fewer data than the randomized complete block design used. Taguchi methods also optimized a robot's process capability.