Challenges in Representing Human-Robot Teams in Combat Simulations

Abstract

Unmanned systems are changing the nature of future warfare. Combat simulations attempt to represent essential elements of warfare to support training, analysis, and testing. While combat simulations have rapidly incorporated representations of unmanned systems into their capabilities, little has been done to distinguish unmanned systems from human systems in these simulations. This is making it difficult to impossible to consider questions of future manned/unmanned system mix, levels of unmanned system autonomy required for most effective operational success, and other relevant questions. One might think that replacing humans with fully autonomous unmanned systems, such as in unmanned convoys, results in identical mission performance with the added benefit of a decrease in loss of human life. However, this is a naïve line of reasoning when one considers that unmanned systems cannot react to the battlespace environment with the same level of flexibility as humans. Unfortunately, we have not yet been able to capture such distinctions in combat models. This paper discusses the challenges we face in developing improved models of human systems, robotic systems, and human-robot teams in combat simulations, with examples posed in the context of the Combined Arms Analysis Tool for the 21^st Century (COMBATXXI), a discrete-event simulation developed and employed by the U.S. Army and U.S. Marine Corps to address analytical questions about future warfighting capabilities.