Crowdsourcing-Based Approbation of Communicative Behaviour Elements on the F-2 Robot: Perception Peculiarities According to Respondents

This article is dedicated to F-2 the companion robot and to interpretations of respondents’ estimations of designed communicative multimodal behaviour. The affective robot is described: it represents a platform for implementing and verifying various individual behavioural traits for robots. F-2 interprets multimodal input: text, face orientation and tactile signals; it translates the input into facts, which are seeds for further affective behaviour. Facts trigger behavioural patterns for reacting—concurrent scenarios with their activation degrees varying over time. The most activated scenario is implemented via one reaction out of a pool of corresponding scenario reactions. Each reaction is multimodal and includes one or several components: speech, gestures, gazes. Robot behaviour is estimated by human assessors during conducted experiments on communication. Several notable effects were observed during perception of implemented communicative behaviour of F-2. These effects are discussed, they are presumed to be evidence to common human expectations transfer from human-human interaction to human-robot interaction.