Abstract
Dull, dirty, and dangerous emergency response tasks (e.g., Fukushima nuclear disaster) typically call for human-robot teams such that the robots are teleoperated by one or more operators at a remote location. However, this distance creates a disconnect between human and robot that hinders effective teamwork. Mixed-initiative interaction has the potential to bridge the disconnect by enabling the human and robot to opportunistically interleave their complementary capabilities. However, allowing robots to intervene and take over from the human could be ill-fated if the intervention is illtimed. Hence, in this article, we examine the question of when should a robot intervene and seize control from a human operator, through a human subject study where the human-robot team is tasked to search for disaster victims in an unknown environment.