Our industries are highly dependent on robots for a wide range of applications such as material transfer, assembly, and welding. Robotic machining application for pre-machining of hard materials or finishing operation with reduced tolerances is a sector which will be growing over the next few years.

In application of curved surfaces machining on large composite supports of 3-4 meters diameter, defects must be limited to a few hundredths. However, many studies1-3 have highlighted inherent limits of articulated robots such as low precision and accuracy under large load representative of rough machining operation. The model gives a representation of the robot, it can be geometric, kinematic or dynamic.

Some research focus on stiffness modeling5. Considering dynamic, flexibility and all robot's elements, these models make possible to obtain a high accuracy. However, cost, complexity and robustness of parameters identification must be considered. To allow a simple and sufficiently relevant industrial use, we choose to study two type of accuracy's corrections.

A Calibration that allows a better knowledge of the constituent elements of the robot by adjusting the geometrical parameters used for its control (centers, lengths and orientation of frames). The Denavit-Hartenberg modified convention proposes an analytic representation of opens structures4. In addition, the Hayati7 model, considers the coaxiality errors between frames of the geometric model. This study is important because geometrical errors represent nearly 80% of positioning static errors8.

The calibration method used in this work based on a linear optimization in sense of least squares which determine corrected values of geometrical parameters to update the model. That accuracy's correction does not consider effects of machining on deviations of the trajectory. This role is dedicated to another type of correction.

Mirror correction method6 makes possible to correct a tool path trajectory based on the error between the measured and the programmed trajectories. This method assumes that deviations of the robot are considered as small displacements. In this article, measurement of cutting forces shows that the measured deviation depends on machining's constraints.

This article describes calibration and mirror correction methods, results highlight a static accuracy gain made by the calibration and improvement of pieces's dimensional quality by the mirror correction.

References:

1 Chen Y, Dong F (2013) Robot machining: recent development and future research issues. Int J Adv Manuf Technol 66(9-12):1489–1497

2 Pan Z, Zhang H, Zhu Z,Wang J (2006) Chatter analysis of robotic machining process. J Mater Process Technol 173(3) :301–309

3 Tobias SA, Fishwick W (1958) The chatter of lathe tools under orthogonal cutting conditions. ASME, Trans 80(5) : 1079–1088

4 W.Kkhalil, E.Dombre (2002). Modeling, Identification & Control of robots. HPS, 11, p. 257-289.

5 Abele, E., M. Weigold, and S. Rothenbücher. “Modeling and Identification of an Industrial Robot for Machining Applications.” CIRP Annals - Manufacturing Technology 56, no. 1 (2007), p. 387-390

6 Dumas, Claire, Aude Boudelier, Stéphane Caro, Sébastien Garnier, Mathieu Ritou, and Benoît Furet. “Développement D'une Cellule Robotisée de Détourage Des Composites.” Mécanique & Industries 12, no. 6 (2011), p. 137-143.

7 S. Hayati and M. Mimirani (1985). Improving the absolute positioning accuracy of robot manipulators. Journal of Robotic Systems, 2(4) p. 397-413

8 Wu, Yier, Alexandr Klimchik, Stéphane Caro, Benoît Furet, and Anatol Pashkevich. “Geometric Calibration of Industrial Robots Using Enhanced Partial Pose Measurements and Design of Experiments.” Robotics and Computer-Integrated Manufacturing 35 (October 2015): 151–68.