Workpiece localization methods for robotic welding - a review

Frequent manual reprogramming of robots is one of the impediments for the installation of robotic welding cells in small and medium-sized enterprises (SMEs). CAD-based offline programming solutions yield potential to reduce these efforts. A central component of these automatic programming approaches is the localization of the workpiece's actual position in the robotic welding cell. This paper presents an extensive literature review on workpiece location for robotic welding as well as latest technological developments gathered at an international trade fair on welding. Comparing the approaches developed for robotic welding with those for other robotic applications such as bin picking, potential improvements are identified. It is reasoned that localization algorithms using offline training and global optimization will increase the flexibility of existing solutions if they are used as feature-based coarse localization independent of any prior knowledge of a workpiece's pose. This lowers the obstacles for SMEs to invest in robotic welding as it holds the potential to reduce setup times in a small lot size production environment.