Measurement of Visitor Behavioral Engagement in Heritage Informal Learning Environments Using Head-Mounted Displays

Measuring visitor engagement in informal learning environments presents critical challenges for optimizing educational experiences and spatial design. While existing research predominantly focuses on formal settings, systematic analysis of multidimensional engagement in complex environments like museums remains underdeveloped. This study introduces the first integrated head-mounted display (HMD)-based framework combining meso-scale spatial analysis through behavioral engagement heatmaps with micro-level temporal engagement modeling via headset pose and eye-tracking data. Our edge-optimized long short-term memory (LSTM) model achieves real-time engagement measurement with a mean squared error (MSE) of 0.145 using four physiologically grounded features from a two-stage user study (N=20 for feature analysis, N=15 for modeling). The framework synthesizes planar trajectory heatmaps and panoramic fixation distributions to enable both real-time adaptive support and post-hoc exhibition design insights. Results demonstrate HMDs' potential as precision measurement tools, establishing methodological foundations for intelligent heritage environments that dynamically respond to engagement states through integrated meso-micro analytics.