Robotic welding accounts for about one fourth of all industrial robotic tasks, yet automation is lacking. A novel low-cost robot with automatic track calculation and quality control will have major impact on quality, costs and competitiveness.
Manual welding is time consuming, and existing automated robots are not
efficient enough at reprogramming to support small batches of products.
Development of flexible and reprogrammable industrial robots is a key
driver of economic success, particularly for the numerous small and
medium-sized enterprises (SMEs) supporting the EU manufacturing sector.
The ambitious EU-funded project 'Decision making and augmented reality support for automatic welding installations' (
MARWIN) addressed the challenge with development of a cognitive computer vision-based welding robot. The simulation environment is the core of the system, calculating the required kinematics for robot movement. It is an offline robot programming solution with intelligent and flexible adjustment tools that can reduce robot programming time and create tailored programmes for a given part based on computer-aided design (CAD) files.
A 3D object reconstruction and registration system reproduces the object with accuracy of about 500 microns and can also find surface discrepancies. It enables comparison of specifications from a CAD file to the actual assembly. Automatic adjustment of robotic and welding parameters to minimise deviations gets a helping hand from a large database of welding parameters and positions.
The tremendous business opportunity for the SMEs involved in the MARWIN project is expected to yield close to EUR 3 million within the first 5 years of commercialisation. Given the importance of welding to the EU economy, the widespread and long-term benefits for Europe are expected to be considerable.