A contribution to robot tooltip spatial accuracy identification

Abstract

A stochastic model and a computer simulation algorithm for combining all measurable k inematic/geometric and misalignment error components are discussed. The aim is to predict the overall quasistatic tooltip spatial positioning accuracy for a gantry robot. The errors include joint position inaccuracies, joint misalignments, joint zero position offsets, imperfect axis directional straightness, squareness errors and kinematic coupling effects. The interaction between robot joints and the coupling between these error components are very complex, making the determination of the overall robot spatial position accuracy difficult. A computer simulation program based on the above model and algorithm has been developed for predicting overall robot spatial position accuracy. Simulation results are compared with actual spatial accuracy test results using a computerized Theodolite system. It is shown that the overall 3-D spatial error histogram for simulation of 1150 random poses in a large gantry robot work envelope is a chi-square distribution.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>