Measurement method for 2-dimensional coordinate system

Abstract

The subject of this Master’s of Science Thesis is to study suitability of predefined technologies and development of a measurement system for measuring Cartesian style robot’s positioning accuracy within its x/y-plane. The methods currently used for measuring and calibrating robot positioning are based on commercial solutions or measurement systems developed internally. Both of these have their non-beneficial properties that lead to the commissioning of this thesis work. The commercial system provides a high accuracy and repeatability performance but are often costly and difficult or impossible to integrate to the robot system in question. The second mentioned measurement system has much lower building costs but it only measures in one axis at a time and thus, it doesn’t see the possible orthogonality errors. This problem was approached with three methodologies covering the theoretical background of the main geometrical error sources, some basics of the theory of measurement uncertainty, evaluation procedure for evaluating suitability of technologies for measuring positioning and developing a measurement system with the selected technology based on the evaluation results. The technologies are evaluated with a scoring system based on criterion that sets a range of requirements for e.g. accuracy performance and other functionality. The goal of this thesis was to develop a low cost, system integrable positioning measurement system and study its performance and usability by comparing it to commercial measurement systems and systems design and developed internally. After conducting evaluation of the technologies, a camera-based technology was selected for development phase. This technology solution includes a camera and optics that are the robots standard equipment and thus won’t add any additional costs. This solution requires a measurement target which is used with the camera to detect motion and algorithms for camera calibration and for calculating motion increments from the captured images. With the combination of camera, optics and the measurement target a value of 55.2 pixels per millimeter was obtained which translates to 17.9 μm ± 0.001-pixel size in the measurement images.