Emotional Contagion System By Perceiving Human Emotion Based on Physiological Signals

摘要

The various wearable equipment enables the robot to measure some people emotional analysis by physiological signals. In our research, we present one emotional contagion system which could develop robotic emotion based on human interactive emotion. First, we research the physiological emotion recognition as the robotic sensory generation. Second, this paper considers the influence of the sensory on perception generation and establishes a tensor representing the real-time perception. Finally, based on previous work, we establish the emotional transition model for generating the robotic emotion and considering the correlation between the previous perception and emotion for cognitive and emotional continuity. In experiment part, we invited six postgraduates wrapping the wearable strap to watch 40 kinds of affective videos. We recorded their physiological signals while be transferring to the robot synchronously. We find emotional contagion system could have the superior performance with whatever is single user's interactive situation or multi-users' one.