Mitigating this challenge is the subject of the industrial PhD project, “Parawood: Framework for On-site Robotic Timber Fabrication”, carried out in close collaboration with the industry partner; Odico Construction Robotics. Odico has developed a transportable robotic fabrication concept, Factory-on-the-Fly (FotF), which this project aims to extend with new processing capabilities and instruction methods. The FotF system is a closed envelope (trailer or container) containing an industrial robot arm and a unique robotic work environment for manufacturing tasks. Instructions for the tasks are given through a software app on a tablet that configures fabrication actions for element(s).
This research is focused towards developing a solution that can assist carpenters’ on-site practices, a construction process commonly carried out on-site. However, equipping carpenters with a robotic solution presents two challenges; 1) how could the robot fabricate elements that support carpenters’ current practice? And 2) how can carpenters, who are usually robotic non-specialists, best instruct fabrication information?
In response to the first challenge, the project seeks to develop a method to re-design the Factory on the fly to process timber. This is achieved through constructing three 1:1 prototypes, of which one was the first commercially robotically fabricated timber house in Denmark that is situated in Odense. Findings from the prototypes has led to a proposal for a semi-autonomous, and possibly in the future transportable, robotic system for making timber structures for houses. This system is currently being used as a production unit at Odico Construction Robotics.
In response to the second challenge, the project has developed a technique that enables to instruct robotic actions through hand drawn notations made directly on work piece. This breaks the current paradigm where robotics are commonly instructed through written code. As part of the developed technique the project has developed a software that relies on Computer Vision and AI – or Machine Learning to develop drawn language to communicate robotic fabrication actions.
The project is an industrial PhD that has been made with Odico Construction Robotics A/S and with funding from the Danish Innovation Fund.
- Associate Professor Claus Peder Pedersen
- Asbjørn Søndergaard, Chief Technology Officer, Odico
- Dagmar Reinhardt, Associate Professor, The School of Architecture, Design and Planning, The University of Sydney