Developing Human-Robot Team Interdependence in a Synthetic Task Environment

Abstract

In future urban search and rescue teams, robots may be expected to conduct cognitive tasks. As the capabilities of robots change, so too will their interdependence with human teammates. Human factors and cognitive engineering are well-positioned to guide the design of autonomy for effective teaming. Previous work in the urban search and rescue synthetic task environment (USAR-STE) used Minecraft, a customizable gaming platform. In this effort, we advanced the USAR-STE by increasing interdependence in dyadic human-robot teaming through the Coactive Design framework. In this framework, we defined required capacities of victim identification in USAR from literature, and used them as inputs for modeling interdependence, and determined recommendations that would enhance interdependence in the task environment. Although Coactive Design is typically used to design interdependence for robots or jobs, we demonstrated how it can also be used to design an experimental team task environment.