Since 2015

The aim of this work is to show the potential of robotic fabrication and computer vision to autonomously assemble non-conventional masonry structures.

Robotic fabrication of Abeille-like curved wall

In our research we build Abeille-like curved wall where all the ashlars can have different shapes. For the technique to be easily and robustly usable on a construction site, we only rely on simple 2D cameras rather than 3D scanners. Pose (location and orientation) estimation of the ashlars according to the robot is performed through the training of a convolutional neural network trained on a dataset of synthetic images.

A masonry structure with ashlars whose shapes are all slightly different could be therefore easily built, opening a new way to think masonry structures.