Adaptive optimization and interaction method for immersive exhibition space experiences integrating multimodal physiological sensing

Interactions in immersive exhibition halls rely heavily on explicit inputs such as controllers and touchscreens, making it difficult to respond in real time to users’ unconscious emotional fluctuations, which frequently disrupts immersion. This paper proposes an adaptive optimization and interaction method that integrates multimodal physiological perception. First, a hierarchical attention fusion network is designed to fuse multimodal physiological signals and achieve high-precision emotion decoding. Second, a deep reinforcement learning control strategy based on the soft actor-critic algorithm is constructed to drive the adaptive optimization of spatial parameters using the decoded emotional states. Finally, a closed-loop implicit interaction chain comprising perception, decoding, decision-making, and response is established. Experimental outcome indicates that the suggested method achieves an emotion recognition accuracy of 93.5%, reduces system response latency to 7.8 ms, and yields a user experience satisfaction score of 9.2. This method provides a feasible technical pathway for building smart exhibition halls capable of emotional empathy.