Adaptive Virtual Environments based on Biofeedback: XR training to improve decision-making and acting in stressful situations
Olilvia Zechner (2025): Adaptive Virtual Environments based on Biofeedback: XR training to improve decision-making and acting in stressful situations Cumulative Dissertation submitted to the faculty of Digital and Analytical Sciences Paris Lodron University of Salzburg in partial fulfilment of the requirements for the Doctoral Degree Dr.Phil.
First responders regularly encounter high-stress situations that demand rapid decision-making and effective action. While simulation-based training is crucial for developing these skills, traditional methods often struggle to replicate the psychological pressures and complexity of real emergencies. This dissertation investigates how extended reality (XR) training environments can be enhanced through the integration of biosensing technology and adaptive scenario control to better prepare first responders for high-stress situations.
The research addresses four central questions: (1) How can immersive technology enhance current first responder training? (2) How can trainee stress be effectively measured and leveraged during XR simulation training? (3) What performance metrics are crucial for enabling effective trainer engagement? (4) What ethical considerations must be addressed when designing adaptive virtual environments for first responder training?
Previous research has demonstrated XR’s potential for safe, repeatable training scenarios, while studies in biosensing have shown promising applications for stress monitoring. However, the integration of these technologies for dynamic, personalized training environments remains largely unexplored, particularly in the context of first responder preparation.
This research aims to bridge this gap by developing and evaluating novel approaches to XR training that incorporate real-time physiological data and performance metrics. The goal is to create more effective, personalized training experiences that better prepare first responders for the psychological and operational demands of their roles.
The methodology combines user-centered design principles with mixed-methods research, including human factor studies, field trials and system evaluations. Data was collected from over 600 participants across multiple law enforcement and medical first responder organizations through workshops, interviews, surveys and practical training sessions.
Key findings include: (1) The successful development of a stress dashboard that enables real-time monitoring of trainee stress levels; (2) Identification of crucial performance metrics for different first responder profiles; (3) Implementation of adaptive scenario control mechanisms that respond to trainee stress and performance; and (4) Development of ethical guidelines and recommendations for implementing AI-driven adaptive training environments.
This research advances the field by demonstrating how biofeedback mechanisms and adaptive training techniques can be effectively integrated into virtual simulation environments. The findings provide practical guidelines for designing more sophisticated and personalized training systems, while also addressing crucial ethical considerations. The developed frameworks and methodologies contribute to both the theoretical understanding and practical implementation of next-generation training solutions for first responders, with potential applications in other high-stress professional training contexts.
