Sub-Millimetric Fiducial Localization for Intralogistics Robotics

This study presents a novel approach to achieve sub-millimetric localization in intralogistic robotic applications by combining the simplicity of ArUco code detection with the enhanced accuracy of pose estimation using a structured light 3D camera. By evaluating the point cloud generated by the depth camera, the cartesian pose of the ArUco code can be directly extracted, eliminating the need for prior knowledge about camera intrinsics or calibration. This method enhances the accuracy of relative robot localization within warehouse environments up to 0.1 mm at a 1 m distance. To the best of our knowledge, this accuracy has not been reported in any previous work in the literature.