Responsive Social Positioning Behaviors for Semi-Autonomous Telepresence Robots

Social interaction with a mobile robot requires the establishment of appropriate social positioning behaviors. Previous work has focused mostly on general and static rules that can be applied to robotics, such as proxemics. How can we deal effectively and efficiently with the dynamic positioning common in social interactions, such as the leaning behaviors we observed in conversations between elderly with hearing problems? We propose to refine the existing approach by having a robot continuously adapt its positioning behavior based on the reactions people give to its earlier actions - i.e. by making the robot responsive to feedback cues. To implement such a responsive system, we will have to develop systems for the detection of these feedback cues, as well as strategies to adapt the robot's behavior based on them.