Multimodal Physiological Sensing for Adaptive Learning Environments

The COVID-19 pandemic has highlighted critical challenges in maintaining student engagement across educational settings, with decreased attendance and participation affecting learning outcomes. The multidimensional nature of student engagement, encompassing emotional, behavioral and cognitive components, presents significant methodological challenges in measurement despite its established role as a crucial predictor of academic resilience. Although computer science research has integrated sensors and cameras for monitoring student states, existing solutions face privacy concerns, environmental dependencies, and scalability issues, with no freely available frameworks for researchers. This research proposes an open-source intelligent interface system that integrates multimodal physiological sensing with real-time logging and analytics to study and enhance teaching effectiveness. By monitoring physiological indicators through a distributed architecture, the system enables classroom-wide monitoring while preserving privacy. The framework utilizes wearable biosensors and ambient audio analysis to create a comprehensive engagement monitoring platform that can record, detect, and respond to the complex interplay between emotional states and learning processes. Beyond its technological contributions, this research aims to enhance educational accessibility and inclusion, particularly for disadvantaged and at-risk populations. The system ability to provide objective and real-time insights supports data-driven teaching adaptations, representing a step toward more effective and adaptive learning environments.