For on-machine measurement of workpiece position, orientation, and geometry on machine tools, five-axis continuous (scanning) measurement by using a laser displacement sensor has a strong advantage in its efficiency, compared to conventional discrete measurement using a touch-triggered contact probe. In any on-machine measurement schemes, major contributors to their measurement uncertainty are error motions of the machine tool itself. This paper formulates the influence of geometric errors of rotary axis average lines on the measurement uncertainty of the five-axis on-machine measurement by using a laser displacement sensor. To validate the present simulator, experimental comparison of measured and simulated trajectories is conducted on five-axis on-machine measurement of a precision sphere of the precalibrated geometry. For total 28 paths measured on the spherical surface, an error in the simulated trajectories from measured trajectories (properly low-pass filtered) was at maximum 5 μm. Uncertainty assessment demonstration for more practical application example of a turbine blade measurement is also presented.
18 March 2022 in IEEE/ASME Transactions on Mechatronics
To improve a robot’s absolute positioning accuracy, researchers have extensively studied the robot kinematic model containing position and orientation errors of rotary axis average lines, widely known as Denavit–Hartenberg (D-H) parameters. To further improve the absolute positioning accuracy of industrial robots, this article proposes a novel kinematic model and its identification scheme. The proposed kinematic model for a serial-linked industrial robot contains the bidirectional angular positioning deviations of each rotary axis, represented in a lookup table, in addition to its D-H parameters. The angular positioning deviations of the rotary axes are modeled as a function of angular command positions, along with the direction of rotation to model the influence of backlash. This article also proposes a novel approach to identify the proposed kinematic model with the bidirectional angular positioning deviations using a laser tracker with indexing each rotary axis at specified angular positions (“circle point method”). Moreover, the model-based compensation technique is being experimentally investigated to validate the prediction accuracy of the proposed model. The findings of the experiment show that the proposed model enhances the robot’s absolute positioning accuracy significantly over the entire workspace.