A new experimental method for predicting static tool displacements for machining with robot manipulators

In this paper, we present a new method for predicting static displacements of the robot when the tool is subject to external forces such as during machining. In particular, we present a new, cheap and simple way of generating and detecting highly varying forces and associated tool deviations aimed at accounting for elastic deformations in the joints and deformations due to joint backlash and other sources. We provide results from initial experiments on a Kuka KR5 R650 robot where we by only studying the elastic part is capable of reducing the error to 22% of the uncompensated errors.