An Interaction Taxonomy of Human–Agent Teaming in Next Generation Combat Vehicle Systems

Abstract

Next Generation Combat Vehicles (NGCVs) are incorporating more advanced technology which will enable humans and intelligent artificial agents to team up on the battlefield. Effective system design and evaluation for these human–agent teams require an understanding of individual and team tasks in the context of larger-scale operations. Previous taxonomies of human–automation interaction and human–agent teaming have been proposed, however, there is a lack of work focused on team interactions in the military domain and the teamwork dynamics required for our purposes are not captured. Unstructured interviews with subject matter experts, manuals, and relevant literature were synthesized, and a task analysis was conducted to develop a novel interaction taxonomy approach consisting of three primary categories, each with multiple dimensions: task, team composition, and communication. This taxonomy may generalize into human–agent teaming within a variety of NGCV crews and serve as a model for characterizing human–agent team interactions in other domains.