Study on adaptation of robot communication strategies in changing situations

Abstract

With the increasing demand for socially assistive robots in various domains, it becomes important to make robots that are able to start communication and attract people’s attention whenever necessary. Although verbal and non-verbal ways of communication have been studied before, adjusting them to changing situations of the environment is still a challenge. In fact, as the environment in which the interaction takes place mutates, and individual differences exist in human perception, some channels of communication may be less effective than others. This paper describes a study in which a Nao robot tries to use different behaviours to start an interaction with a human counterpart, in four changing environmental conditions (loudness, ambient luminance, distance and angle). The robot determines the state of those conditions and through an online learning process is capable of choosing the most effective behaviour for each individual taking into account individual preferences. The findings of this paper will be useful to make a robot initiate a communication successfully.