Task scheduling for human-robot collaboration with uncertain duration of tasks: a fuzzy approach

Collaborative robotics is a new production paradigm for manufacturing systems. Typical applications are collaborative assemblies, requiring humans and robots to share a common space to accomplish common tasks. This imposes to predict the actions of humans and consequently control the robot in a safe way, making the time required for completing a task non deterministic as the robot could be forced to be slowed down in unpredictable ways. Moreover, the uncontrollable nature of the human mate as well as the intrinsic variability of humans completion times introduce other sources of uncertainty to account for. This work aims at proposing a novel fuzzy scheduling approach for managing in an optimal way the uncertainties arising in collaborative assemblies. Realistic experiments, involving the assembly of two products, are performed using the dual arm robot YuMi of ABB.