Communication atmosphere in humans and robots interaction based on fuzzy analytical hierarchy process

Abstract

Communication atmosphere in Human-Robot Interaction (HRI) is estimated by integrating emotional states of humans and robots based on the concept of Fuzzy Atmosfield (FA), where human emotion is estimated from bimodal communication cues (i.e., speech and facial expression) and robot emotion is generated by emotional expression synthesis. Fuzzy Analytical Hierarchy Process (FAHP) is used for dynamic weight calculation of emotional states of humans and robots in the FA model, in which a three-level hierarchy is adopted for the analysis of communication atmosphere and four acoustic cues (i.e., volume, rate, pitch, and duration) are utilized to calculate the dynamic weights. Experiment is conducted in a multi-modal emotional communication based humans-robots interaction (MEC-HRI) system, by which experimental results demonstrate the validity of proposed FA model. Understanding the communication atmosphere can help to facilitate smooth communication between humans and robots by adjusting robot behavior for emotion expression. The proposal is being planned for an atmosphere representation system for realizing causal communication in HRI and background music is being considered for the FA model.