|This project is funded by European Robotics Challenge (EuRoC)
Challenge 1. Reconfigurable Interactive Manufacturing.
The feeding of parts from bulk supply to production lines is a common but still challenging task in industrial automation. We present an industrial prototype that has two objectives: first, it needs to be easy reconfigurable to handle new part variants, and second, it should be safe for collaboration with human workers. We propose a reconfigurable vision system that finds generic graspable parts and robustly estimates 6D pose given CAD data with a learning method based on synthetic data. To guarantee that the system is human safe, we use a cable-driven lightweight igus-robolink® that can move at high velocities without risk of harming human workers. We improve the safety by detecting persons entering the workspace of the arm and quickly stopping its movement.
Picking test (16/19 success)
Feeding test (9/10 success)
Seongyong Koo, Grzegorz Ficht, Germán Martín García, Dmytro
Pavlichenko, Martin Raak, and Sven Behnke:
Robolink Feeder: Reconfigurable Bin-Picking and Feeding with a Lightweight Cable-Driven Manipulator
Accepted for 13th IEEE International Conference on Automation Science and Engineering (CASE), Xi'an, China, to appear August 2017.