Event Location: NSH 3305

Abstract: Human conversations are a type of jointly focused encounter in which the participants gather together and openly cooperate to sustain a single focus of attention. When the participants are standing, their cooperation can be observed in their spatial organization. People position and orient themselves to maximize their opportunity to monitor one another’s mutual perceivings during conversations.

The goal of this dissertation is to enable robots to reason about the spatial patterns of behavior that emerge during free-standing conversations in order to detect these jointly focused encounters and cooperate to sustain them. To this end, we first develop a flexible robotic platform for studying human-robot interactions, and conduct experiments to examine various aspects of group conversations involving robots. Second, we propose a general framework to track the lower-body orientation of the people in a space and detect their conversational groups based on their spatial behavior. This framework takes advantage of the mutual dependency between these problems. People’s lower-body orientations are key for describing spatial arrangements that we can use to detect free-standing group conversations. At the same time, the location of conversational groups in a space can bias the interactants’ lower-body orientations as well as that of nearby people. Third, we propose to reason about group membership uncertainty during motion planning. We will develop a mechanism to decide if and when a robot should reconfigure during conversations to improve group membership estimation.

The proposed methods will be validated on data captured from our platform and during real-time human-robot interactions. For these evaluations, we developed new experimental protocols to study group conversations with robots. These protocols are easy to adapt, and their success will demonstrate their value to future multi-party HRI explorations.

Committee:Aaron Steinfeld, Co-chair

Scott E. Hudson, Co-chair

Kris Kitani

Brian Scassellati, Yale University