Robust assembly task assignment in Human Robot Collaboration as a Markov Decision Process problem

Abstract

Collaborative robots can work together with human workers in assembly workstations. Their drawback is the lack of flexibility that force human co-worker to bear the cognitive burden of strictly replicating every time the same tasks. To improve human-robot collaboration, human should be allowed to exchange tasks with the robot if this doesn't hinder the final assembly. The study proposes a robust real time optimization of the assembly task assignment through the modelling of the assignment problem as a Markov Decision Process with a randomly selected starting state.