Analytical basis for evaluating the effect of unplanned interventions on the effectiveness of a human-robot system

摘要

Increasing prevalence of human-robot systems in a variety of applications raises the question of how to design these systems to best leverage the capabilities of humans and robots. In this paper, we address the relationships between reliability, productivity, and risk to humans from human-robot systems operating in a hostile environment. Objectives for maximizing the effectiveness of a human-robot system are presented, which capture these coupled relationships, and reliability parameters are proposed to characterize unplanned interventions between a human and robot. The reliability metrics defined here take on an expanded meaning in which the underlying concept of failure in traditional reliability analysis is replaced by the notion of intervention. In the context of human-robotic systems, an intervention is not only driven by component failures, but includes many other factors that can make a robotic agent to request or a human agent to provide intervention, as we argue in this paper. The effect of unplanned interventions on the effectiveness of human-robot systems is then investigated analytically using traditional reliability analysis. Finally, we discuss the implications of these analytical trends on the design and evaluation of human-robot systems.