Uncertainty evaluation in robot calibration by Monte Carlo method

In this paper it is presented a technique to evaluate the calibration uncertainty for a robot arm calibrated by circle point analysis method. The method developed, based on probability distribution propagation calculation recommended by the Guide to the Expression of Uncertainty of Measurement, and on Monte Carlo method, makes it possible to calculate uncertainty in the identification of each robot single parameter, and thus to estimate robot positioning uncertainty in accordance to its calibration uncertainty, and not according to a set of single locations and orientations previously defined for a unique set of identified parameters. Besides, this technique allows beforehand to establish the best possible conditions for the data capture test for the identification, which turn out to have the less possible calibration uncertainty, according to the variables involved in the data capture process for the identification, by propagating their influence up to final robot accuracy. Currently, the results validit...