“Predicting Performance Times of Robot Systems”

Abstract

It is essential that for efficient design of manufacturing systems, reliable techniques are available for predicting the performance of mechanical handling equipment including robots. Because of the analogies frequently drawn between robot and human operator anthropomorphism(1), it is logical that one approach to performance prediction should be via Robot Time and Motion (RTM) studies(2). This leads to the concept of a computer based RTM analyser, the essential components being;

• (a) sets of element tables for given robot designs

• (b) computational performance models for given robot designs

The task input to the RTM analyser is the RTM specification of work method for a given industrial robot, whilst the output is the performance prediction for the specified method. Reference 2 uses a linear regression equation to predict elemental RTM times as a function of the distance travelled during a specific motion. However, the reference also indicates that there is a surprisingly large variation about the mean trend line when measurements are made on the Stanford Arm. It is also possible to show that the approach can lead to large prediction errors when the robot motion pattern contains a number of small segments(3).