Repeatability Assessment of Industrial Robots According to ISO 9283 Using a Vision-Based Measurement System

This thesis presents a detailed evaluation of the repeatability performance of the RTA ARTIC6- 1215-6kg industrial robot, conducted in compliance with ISO 9283, the internationally recognized standard for testing industrial manipulator performance. Repeatability—defined as a robot’s capacity to return consistently to a predefined position over repeated cycles—is a critical metric for ensuring precision and reliability in industrial automation tasks. The study was structured in two main experimental phases, each employing a different measurement approach. In Phase 1, a conventional contact-based method was used, involving a 3D-printed spherical end-effector mounted on the robot. Measurements were recorded using a three-axis mechanical dial indicator fixture with orthogonally arranged indicators. Thirty repetitions of a programmed movement between two target positions (P1 and P2) were performed to form the statistical foundation for the analysis. Phase 2 employed a non-contact measurement approach using a calibrated, high-resolution optoelectronic Optitrack system. This vision-based setup included two subtests: (a) Phase 2a, which repeated the P1 to P2 motion while carrying a 4 kg payload to assess the impact of load, and (b) Phase 2b, which evaluated repeatability at five spatial positions arranged in an ISO cube configuration. These multi-point measurements offered insight into the robot’s consistency across its working envelope. Trajectory planning and robot control were carried out using Motion Perfect software, while RoboDK was used for offline simulation and collision checking. MATLAB was employed for data processing, sphere center estimation, and statistical computations. Microsoft Excel was used for result visualization and table preparation. All mechanical components, including the 20×20×20 mm calibration cube and the end-effector, were designed in Autodesk Inventor and fabricated using additive manufacturing. Results from Phase 1 demonstrated excellent repeatability under unloaded, contact-based conditions, with a positional deviation of just 0.0260 mm. In Phase 2a, the presence of the payload led to a slightly higher deviation of 0.0406 mm. Phase 2b showed consistent results across most points, though one position (P5) exhibited significantly greater variation (up to 0.2628 mm), highlighting the need to assess repeatability across the full workspace. This work emphasizes the value of integrating traditional and modern metrology methods for comprehensive robot performance evaluation. It provides practical insights into the spatial consistency of robotic motion and underlines the importance of standardized testing protocols for use in research, integration, and industrial deployment. Keywords: Industrial robotics, repeatability, ISO 9283, mechanical dial indicator fixture, vision-based measurement, Optitrack system, P1–P2 trajectory, ISO cube configuration, Autodesk Inventor, MATLAB, RoboDK, Motion Perfect, robotic performance testing.