Recycling decommissioned large metal structures, e.g. oil rigs and ships, or equipment, e.g. large engines requires them to be dismantled, moved to a metal scrap yard and cut into small workable chunks. Currently, the cutting operation is conducted manually by skilled workers using a gas torch. This manual operation is slow and laborious. Due to the variety of the scrap pieces and the difficult and unstructured nature of this process, automating this task has many challenges: For each piece, the cutting locations and trajectories need to be determined, the cutting parameters need to be identified, and the cut needs to be executed at certain torch speed and poses. All these operation variables are successfully determined and applied by skilled workers, but very challenging to translate into robot task parameters.

Manual metal scrap cutting

We propose a novel framework for robotic metal scrap cutting in unstructured scrap yards. The cutting trajectories are identified and extracted by the robot using an active vision system. This is the first step towards a robotics cutting scheme in unstructured environments.


Our papers related to robotic metal scrap cutting:

Towards Robotic Metal Scrap Cutting: A Novel Workflow and Pipeline for Cutting Path Generation,
J. Akl, F. Alladkani, B. Calli
IEEE International Conference on Automation Science and Engineering (CASE), 2021.
[Paper] [Video]