Design and Implementation of a comprehensible cognitive assembly system

Abstract

Assembly relies strongly on human factors in high wage countries. It needs to integrate collaborative components in today’s production systems, which are required to perform various cooperative tasks with multiple entities rather than monolithic solutions. Modular concepts introduce a high amount of planning and complexity in volatile environments with unforeseen requirements. Self-optimization is mainly addressed in the virtual domain, today. Economic assembly however needs versatility on both the virtual planning and the real assembly side. Reconfigurable multi robot systems with high absolute pose accuracy and synchronization are extended by large-volume measurement systems as well as local sensors to provide virtually aligned state estimations of multiple end effectors in a global reference system. Assembly systems are then enabled to navigate in the virtually planned machine state by socio-technical realizations of self-organization; they autonomously form a complete self-optimizing assembly. Implementations are shown for cooperating robots and a collaborative process for alignment of fuselage structures.