Automated Analysis of Proxemic Behavior: Leveraging Metrics from the Social Sciences

We discuss a set of metrics for analyzing human spatial behavior motivated by work in the social sciences; specifically, we investigate individual, attentional, interpersonal, physiological, and organizational factors that contribute to social spacing (proxemics). We then present a pilot study to demonstrate the feasibility of real-time annotation of these spatial behaviors in multi-person social encounters. In particular, we are interested in working with sensor suites that (1) are non-invasive to participants, (2) are readily deployable in a variety of environments—ranging from an instrumented workspace to a mobile robot platform—and (3) do not interfere with the social interaction itself. Finally, we provide an analysis of the performance of the annotation system, and discuss applications and future directions of this work. Keywords-Proxemics; spatial interaction; spatial dynamics; spatial orientation; human-robot interaction; PrimeSensor; Kinect.