2018 · 17 citations · 13 references
EngineeringManufacturing EngineeringRobotics EngineeringAutomationIndustrial RoboticsRoboticsPose EstimationRobot CalibrationRobot ManipulationRobot ManipulatorsRobotics In ManufacturingSensor CalibrationAccuracyCalibration Method ValidationFinal SolutionGround Truth AlignmentManipulator CalibrationOptimization-based CalibrationRobotic CalibrationWorkpiece Pose Estimation
This paper presents a method for maintaining high accuracy of laser-guided robotic manipulation by continuously compensating for position and alignment errors for aerospace manufacturing processes. The objectives of the method are to use the robot to perform reconfiguration in a flexible fixture and minimize the calibration work (tool, workpiece), continuously track the position and alignment of the mounted tool on the robotic arm, and minimize the positioning and alignment errors between the tool and the workpiece. The proposed method basically estimates first the transformation matrix between the 6 Degrees-of-Freedom (DoF) reflector sensor, which is mounted on the robotic arm with respect to robotic flange, and second the transformation matrix between the laser tracker frame with respect to the robot base frame. For estimating the two matrices, two solvers based on an iterative absolute orientation method and hybrid optimization solver are presented and thoroughly discussed. The calibration method is first validated in a robotic simulation environment to visualize the transformations before (initial guess) and after (final solution) calibration. The results show that the two calibration solvers are able to detect the exact poses (ground truth) from a simulated data set. Second, the method is applied to a laser assisted robotic system which is composed of Leica Absolute Tracker AT960 and the reflector Leica Tracker-Machine Control (T-Mac) as well as a 6 DoF KUKA KR 60 robotic arm. The method is used to successfully register the laser measurements with respect to the robot base frame.
13