Non-Dyadic Collaboration In Human-Robot Interaction

With the ever-increasing number of domains in which we encounter robots - be it in industry, airports, or the home - the opportunity to interact and collaborate with these grows. And while an abundance of Human-Robot Interaction (HRI) literature has investigated dyadic interaction, non-dyadic HRI research, i.e., more than one human and one robot, has just recently begun to receive increasing attention. In this dissertation, I investigate characteristics of non-dyadic Human-Robot Interaction and collaboration. Specifically, I investigate two research questions focusing on i) the identification of existing characteristics of non-dyadic Human-Robot Interaction research and ii) the influence robots have on non-dyadic collaborative efforts.

This dissertation's contribution is based on five research papers. Paper I presents an empirical investigation of existing research on non-dyadic HRI over the last 15 years. Paper II to IV present qualitative field studies in the domestic and industrial contexts. Lastly, Paper V presents a mixed-methods lab-based study investigating human group collaboration and identifies design considerations to improve non-dyadic human-robot collaboration. Based on these five papers, this dissertation presents two primary contributions.

Firstly, I identify characteristics of non-dyadic HRI through an investigation of 164 research papers. These characteristics include the ongoing paradigm shift from a dyadic focus towards a non-dyadic focus, three non-dyadic configurations within HRI (one-to-many, many-to-one, and many-to-many) and an imbalance emphasising research involving one human interacting with multiple digital artefacts (one-to-many), a classification framework for non-dyadic Human-Robot Interaction, as well as empirical evidence showing the focus of non-dyadic HRI research on simultaneous over sequential interaction.

Secondly, I present several ways in which robots influence collaboration during non-dyadic Human-Robot Interaction. I highlight how introducing robots in both the domestic and industrial contexts into non-dyadic settings can lead to a fragmentation of previously coherent tasks while only some of the sub-tasks are automated. Furthermore, I show how the robot's presence, as previously hypothesised---can lead to a spatial restructuring resulting in a positive change in interpersonal relationships amongst collaborators. Lastly, I argue for the robot's capacity to alter, remove, and create roles and responsibilities within the non-dyadic collaborative Human-Robot Interaction.

Future work includes the investigation of i) robots as pro-active collaborators, ii) increase of transparency during robot introduction to counter unintended negative side-effects, and iii) a reconsideration of what a collaborative robot and collaboration with robots means.