Human-robot collaboration in manufacturing: Quantitative evaluation of predictable, convergent joint action

Abstract

New industrial robotic systems that operate in the same physical space as people highlight the emerging need for robots that can integrate seamlessly into human group dynamics. In this paper we build on our prior investigation, which evaluates the convergence of a robot computational teaming model and a human teammate's mental model, by computing the entropy rate of the Markov chain. We present and analyze the six out of thirty-six human trials where the human participant switched execution strategies while working with the robot. We conduct a post-hoc analysis of this dataset and show that the entropy rate appears to be sensitive to changes in the human strategy and reflects the resulting increase in uncertainty about the human next actions. We propose that these results provide first support that entropy rate may be used as a component of dynamic risk assessment, to generate risk-aware robot motions and action selections.