A network science approach to future human-robot interaction

Abstract

The vision for future Soldier-robot relationships has supported the transition of the robot's role from a tool to an integrated team member. This vision has provided support for the advancement of robot autonomy and intelligence as a means to better support action and cognitive decision-making in the network-centric operational environment. To accomplish this goal, the Soldier's perspective of the human-robot interaction must be further developed, as it directly impacts overall situation management: mission planning, operational roles, function allocation, and decision-making. Here we present a theoretical concept paper that promotes using the foundation of network science to better understand how and why advances in effective Soldier-robot situation management may be realized. We begin by providing a primer on how a network science approach may be used to understand multi-agent teams and network-centric operations. This is followed with a review on the impact of human perception on the human-robot team network structure. Two key points are highlighted. First, the network structure is influenced by the extent to which a Soldier-robot coupling performs independent operations. Second, the degree of automaticity for several properties of the robot specifies the strength of their networked relationship. We conclude with possible advantages of using a network science approach for understanding situation management of Soldier-robot teams in an operational environment. This approach provides a structure for creating visual maps of team structures to understand perceived and anticipated role interdependency, which thus provides the foundation for developing a mathematical description of the dynamic Soldier-robot relationship.