Automating Behavior Selection for Affective Telepresence Robot

The tabletop robot Haru, used for affective telepresence research, enables a teleoperator to communicate affects from a distance. The robot’s expressiveness offers myriad ways of communicating affects through the execution of emotive routines. The teleoperator reacts to input modalities such as the user’s facial expression, gestures and speech-based intent as perceived by the robot’s perception system. However, due to the sheer number of routines to select from, the task of choosing the appropriate or the most preferred routine is becoming cumbersome. In this paper, we propose a human-in-the-loop reinforcement learning mechanism in which an agent learns the teleoperator’s selection preference as a function of the input modalities and aids the routine selection process by narrowing it to n-best optimal choices. Our experimental results show that with only a few number of interactions from the teleoperator, the system can learn to recommend optimal routine behaviors for all perceived modalities, which greatly reduces the workload of the teleoperator.

I. Introduction

A social robot Haru presents a holistic robotic platform that aims to support the study of social presence as well as affective engagement for long-term human-robot inter-action across different contexts [1]. Affective telepresence research has been explored in different studies [2], [3], [4]. Research works with Haru is conducted through a consortium partnership and one of the primary areas that is being actively studied with consortium partners [5] relates to a multimodal robot-human interaction where Haru is used as a medium for affective telepresence from the perspective of the teleoperator. One of the research themes with Haru is to develop interaction capabilities that would facilitate creation a new form of telepresence embodiment that is different from the trivial notion of a tablet-on-wheels and explore techniques for the robot to communicate with people more directly by means of affective computing .