Toward Initiative Decision-Making for Distributed Human-Robot Teams

Abstract

The brief history of human-robot teams can be traced through the changing perspective of a robot's role within the team, which has evolved from being treated as a tool to a recent shift toward the desire to have the robot act as an equal partner. While researchers have made tremendous strides in recent years, "making robots into team players" [1] that can work with humans as peers still presents a multitude of challenges. One key characteristic of a synergistic team is the ability to intervene or backup each other as necessary (e.g., when the other is underperforming). Hence, in this article, we formulate distributed human-robot teamwork in the framework of mixed-initiative interaction, which is an interaction strategy that lets the best-suited member of the team to perform the work by allowing team members to interleave their contributions to the overall team performance through opportunistic seizure and relinquishment of task initiatives. Specifically, this paper aims to address the issue of initiative decision-making - that is, when should a robot take over (relinquish) control from (to) a human teammate.