Manual assembly is error-prone, especially in high-mix, low-volume production. Errors and error prevention create costs - costs that we can reduced by implementing deep-learning based automated quality control. We develop smart cameras that can detect manual assembly steps during live production, allowing for automated documentation, feedback on realised steps and instructions on upcoming steps. Our smart cameras are developed as add-ons to existing digital worker guidance systems, widening their application range.
Part of our team is Jonas Conrad, who is currently finalizing his PhD, where he focussed on researching deep-learning applications in manual assembly.
Felix Schnarrenberger is finalizing his Master in robotics, training in machine learning for computer vision applications.