Data for "Estimating Uncertainty in Robot Kinematics and Pose Measurements with Expectation-Maximization"

Included here are figures and relevant data for the work "Estimating Uncertainty in Robot Kinematics and Pose Measurements with Expectation-Maximization". We present a method to validate the measurement uncertainty of a metrology instrument without a priori estimates in the context of a kinematic calibration using Expectation-Maximization methods and extend our results to characterize post-calibration pose uncertainty for the manipulator throughout a workspace. This technique permits the robot kinematic model to be fitted simultaneously with a parameterized uncertainty model derived from direct-drive laser tracker kinematics. We demonstrate the performance of this algorithm in a simulated and experimental setting, achieving 6.4um position and 70.8 urad rotation error for kinematic calibration and statistically validating the fitted uncertainty model for points throughout the calibrated workspace.