Psychological and Emotional Driven Virtual Reality Human-Computer Interaction Based on Recurrent Neural Network

In the field of virtual reality human-computer interaction, traditional interaction methods are difficult to accurately respond to users' emotional states, resulting in a less natural and immersive experience. To address this issue, this article investigates a psychological emotion driven interaction method based on recurrent neural networks (RNN), aiming to enhance the emotion recognition and response capabilities of virtual reality systems. The article collects physiological and behavioral data of users in virtual environments, including heart rate, facial expressions, and movement trajectories; then, an emotion state analysis model is constructed using RNN, and the model parameters are optimized through training the dataset to improve the accuracy of emotion recognition. The experimental results show that the accuracy of this method in identifying psychological emotions reaches 92%, significantly higher than the 78% of traditional methods; meanwhile, user interaction satisfaction increases by 17.7%. The conclusion points out that emotion driven methods based on RNN significantly enhance the naturalness and user experience of virtual reality interaction, providing a new path for the future development of human-computer interaction.