Loading [MathJax]/extensions/MathMenu.js
Robotic Artistic Writing—Reproduction of Signs Drawn on the Surface of Plastically Deformable Materials | IEEE Journals & Magazine | IEEE Xplore

Robotic Artistic Writing—Reproduction of Signs Drawn on the Surface of Plastically Deformable Materials


Abstract:

Existing surface shaping methods focus on hard materials with stable physical properties. This means that the approaches developed are insufficient for shaping soft mater...Show More

Abstract:

Existing surface shaping methods focus on hard materials with stable physical properties. This means that the approaches developed are insufficient for shaping soft materials. The article describes a cheap method of reproducing plastic deformations using a robotic manipulator. The shape recording and reproduction system consists of two RGB-D cameras, two containers with kinetic sand, and a manipulator with 6 degrees of freedom (DOF). Volunteers create templates by drawing on the surface of kinetic sand with a wooden stylus or finger. Some pressure is exerted while writing. The resulting shapes recorded by the RGB-D camera have the form of ribbon ditches (grooves). The obtained point cloud is processed to create a sand deformation model. In the first stage, a dedicated local smoothing technique is used. Then, special algorithms are implemented to create a description of the main curvatures and key dimensions of recorded signs. A spline-based approach is used. The method allows for the representation of various shapes in a unified form. In the final stage, modulated sinusoidal functions define the robot’s trajectory. The effects of the robot’s operation are recorded to assess the reproduction quality. The point cloud structural similarity measure (Point SSIM) evaluates the results. Experimental research takes into account many different shapes. Copies of shapes created by humans and robots are compared with the originals. The outcomes show that the quality of reproduction achieved by humans and robots is comparable. The median-based curvature similarity measure obtained for the human was only 1.92% higher than the robot’s result, and the covariance-based geometric similarity measure was only 0.74% higher than the robot’s score. The system can be used to mass-produce souvenirs or special implants.
Published in: IEEE Sensors Journal ( Volume: 25, Issue: 10, 15 May 2025)
Page(s): 18090 - 18105
Date of Publication: 03 April 2025

ISSN Information:

Funding Agency:


Contact IEEE to Subscribe

References

References is not available for this document.