Abstract

Emotional interactions in traditional online education often face problems such as unnaturalness, lack of personalization, and neglect of body language. This paper aims to optimize the emotional expression of virtual teachers from the perspective of kinematics and mechanics through the principles of biomechanics, improve the naturalness and personalization of emotional interaction, and thus enhance learners’ emotional involvement, learning motivation, and learning effects. This paper combines the principles of biomechanics to optimize the human-computer emotional interaction system and enhance the emotional resonance between virtual teachers and students. In the study, inverse kinematics and dynamic models are constructed to ensure that the virtual teachers’ movements conform to the laws of human biomechanics and effectively express emotions. Secondly, the facial action coding system is used to model the facial expressions of the virtual teachers, and the coordination of facial expressions and body movements is achieved through a coordinated control algorithm. Finally, an emotion perception and feedback mechanism is designed to enable the virtual teachers to adjust their posture, speech, expression, etc., in real time according to the students’ emotional state and provide personalized emotional response. The experimental results show that the optimized virtual teacher emotional interaction system is significantly superior to the traditional education system in terms of human-computer interaction quality, emotional feedback, and learning motivation. Specific data shows that the experimental group scores 4.3 in positive emotions (positive affect, PA), significantly higher than the control group’s 3.1. In terms of pleasure scores, the experimental group scores 4.5, while the control group only scores 3.2. In addition, the experimental group is significantly better than the control group in various indicators of learning motivation, and its learning time is significantly longer than that of the control group. Its task completion and number of interactions are also better than those of the control group.