ANALYSIS OF THE CAUSAL STRUCTURES LINKING PROCESS VARIABLES TO ROBOT REPEATABILITY AND ACCURACY

This article presents the results of an experimental investigation of how process variables affect robot performance in assembly operations. A linear structural relations (LISREL) model is developed to investig ate the causal effects of load, speed, and distance traversed by a rob ot's gripper on the accuracy and repeatability of the robot's performa nce. The model analytically distinguishes between the direct and indir ect effects that speed, weight, and distance traveled have on both acc uracy and repeatability. Reciprocal effects among the performance crit eria highlight the complexity of the task confronting those attempting to assess robot performance. They also demonstrate the inherent inade quacy of combining several individually derived bivariate assessments of a robot's performance to predict how well the robot will perform wh en confronted with a particular complex task. The excellent fit of the model demonstrates that the performance implications of altering task characteristics, and the complex interrelations among the multiple pe rformance measures themselves can be effectively encapsulated by a LIS REL model.