Dynamic task classification and assignment for the management of human-robot collaborative teams in workcells

The rise of interest in collaborative robotic cells for assembly or manufacturing has been attested by their inclusion among the enabling technologies of Industry 4.0. In collaborative cells, robots work side by side with human operators allowing to address a larger production scope characterized by medium production volumes and significant product variability. Despite the advances in research and the availability of suitable industrial robot models, several open problems still exist, due to the shift in the way of working: correct assessment of the economic profitability, definition of a suitable process plan, task assignment to humans and robots, intuitive and fast robot programming. This paper addresses the task assignment problem by proposing a method for the classification of tasks starting from the hierarchical decomposition of production activities. Task classification is employed for workload distribution and detailed activity planning. The method relays on the assumption that tasks should be allocated, exploiting the different skills and assets of humans and robots, regardless of workload balancing. The proposed method was firstly tested on a simplified assembly process executed in laboratory, then it has been applied to the redesign of an actual industrial process.