abstract
- Digital Twin technology has been increasingly adopted in engineering education to replicate authentic laboratory environments and support interactive learning experiences. However, limited evidence is available regarding its impact on students physiological and cognitive states. In this study, a biometric recognition system was implemented to compare the cognitive emotional engagement of graduate engineering students in two instructional conditions: a traditional physical control systems laboratory and a Digital Twin laboratory simulating the same tasks. Multimodal data were collected through portable EEG and physiological sensing devices to monitor alertness, concentration, fatigue, and stress. Machine learning models were trained and validated, achieving accuracies above 92 percent, and were applied to more than 1 3 0, 0 0 0 biometric data samples recorded during both sessions. The results indicate that concentration levels remained similarly high across environments, while physiological alertness was lower in the Digital Twin condition. A stronger correlation between stress and alertness was also identified in the simulated environment. These findings suggest that Digital Twin laboratories can sustain high cognitive engagement but may not fully reproduce the physiological activation associated with physical environments. This study supports the use of multimodal biometrics as a complementary approach for evaluating learning quality in engineering education.