Recognition of emotions and non-verbal communication characteristics in human-robot interactions - the experience of a systematic review

The aim of this review is to analyse previous research focusing on the non-verbal communication patterns of human-robot interaction, and related study characteristics, HRI-related specificities, methodologic background and main findings. The field of robotics is increasingly recognizing the significance of incorporating affective expression and social interaction. Many non-anthropomorphic mobile robots are currently utilised in various fields. However, these robots lack the ability to project facial expressions and cannot be modified to support affective expressions explicitly, which presents a significant challenge in terms of facilitating naturalistic human-robot interaction with appearance-constrained robots. The current paper introduces the results of a systematic review that follows the PRISMA guidelines. The EBSCO Discovery Service Search Engine was used for a systematic search, yielding 114 records. By a multiple-step analysis, seven studies met the criteria. Research indicates that the physical gestures of robots can be as identifiable as those of humans, highlighting the importance of nonverbal communication in human-robot interaction (HRI). Certain personality traits of the participants appeared to significantly influence the recognition of emotions, particularly evident in terms of confidence and valence. The findings suggest that robots should not only be designed to express emotions clearly but also to enhance these expressions in ways that boost users' confidence in their understanding. Improving the quality of emotional expression and non-verbal cues in human-robot interaction may play a crucial role in Industry 5.0. Future research should focus on the identified underrepresented themes to deepen understanding of emotion recognition in HRI.