Human-robot spatial interaction in a hallway

Abstract

In general, previous research in human-robot interaction has provided evidence for the relevance of robotic motion behavior on positive impressions towards robots. In particular, a socially appropriate designed robotic distance behavior, avoiding personal space violations, has been reported to be essential for a comfortable human-robot spatial interaction. Unfortunately, human-robot proxemics lacks systematic explorations of accepted frontal approach and lateral passing distances in a hallway, and existing studies are complicated to relate to each other due to the application of many different robots and a high variance of further incomparable study details. Thus, the primary research goal of this dissertation was to explore accepted frontal approach and lateral passing distances for an autonomously moving robot toward a standing human in a hallway, and supplementary, the relation of these distance preferences to the robot’s level of speed and human likeness. In addition, the present work intends to explore the relevance of a robot’s proxemic behavior for its overall motion acceptance, and people’s underlying psychological motives guiding their preferences. Towards an exploration of these research objectives, a method comprising a series of three studies was conceptualized and conducted in comparable hallway-like settings. In a first study, participants were instructed to actively control a robot’s frontal approach and its lateral passing distance according to their arousing feelings of discomfort. In two successive studies, the analyzed thresholds of comfort were validated for two diverse robots, autonomously maintaining frontal approach and lateral passing distances in a hallway. In addition, potential influences of a robot’s level of speed (0.6m/s, 0.8m/s) and human likeness (machine-, human-like) on subjects’ distance preferences were explored. In these studies, participants’ sensations were captured by using a set of questionnaires. Obtained results uncovered that accepted frontal approach (approximate 0.8m) and lateral passing mean distances (approximate 0.4m) towards autonomously moving robots in a hallway exist, and that these distances were not significantly affected by the simulated autonomy of the robots. However, a faster robot speed (0.8m/s) significantly increased subjects’ distance preferences. The different outward robot designs of the employed robots had no significant influence on subjects’ frontal distance preferences. In contrast, the manipulated level of human likeness showed that a more human-like robot design had only significant influences on subjects’ lateral distance preferences when a real human face was projected on a robot’s screen (decrease in lateral distances by more than 0.1m). Essentially, in line with previous research, the present work shows that a socially acceptable moving robot should consider people’s distance preferences in order to avoid feelings of discomfort, and to increase the perceived safety and the overall motion acceptance. In consensus with prevailing psychological motives in human-human spatial interactions, the present work suggests similar mechanisms in a human-robot spatial interaction guiding subjects’ preferences. Ultimately, beyond the attained mean distances, frontal approach distances of 1.1m and lateral passing distances of 0.6m were suggested for designing a socially appropriate first contact for a large majority of individuals, which in turn supports a facilitated societal integration. Limitations of this research project are discussed and suggestions for future investigations are presented.