Intelligent task repeatability of an industrial robotic arm

Industrial robots are increasingly being used to perform intelligent tasks such as precision assembly, machine control interactions and switching system activations. The ability of a robot to repeat a certain task indefinitely with minimal error is a measure of its repeatability and is a key determinant of its performance in a given task. The goal of this research is to devise a unified formula to determine the repeatability of a 'Lab Volt 520' Robotic Arm in an experiment designed to simulate keyboard-entry tasks. The task is used as a surrogate for intelligent task functionality in general. The robot arm is programmed to type a predetermined character string by using its servo-controlled motion to reach the target key. The main obstacle in such an operation is the deviation of the acquired target from the expected target. This value of Deviation is seen to vary according to three factors, namely Speed of movement, Complexity of motion and Length of Experiment. Sixteen experiments, each with a 2-level factorial were designed and run in random order. Task error was measured by the percentage of characters missed. We analyze the errors occurring with respect to the combined variation of these three effects. The Standard Deviation of the acquired target serves as a tentative response factor for repeatability of the arm. Eventually, the Analysis of Variances (ANOVA) of these values provides us with noise-less coefficients against each factor and their combinations.