Electrodermal Insights into Stress Dynamics of AR-Assisted Safety Warnings in Virtual Roadway Work Zone Environments

2025-05-15

AI Summary

This study investigates the dynamic impact of an augmented reality (AR)-enhanced multisensory warning system on workers’ stress levels within virtual roadwork zones. Method: Leveraging a high-fidelity virtual reality (VR) construction environment, we integrated AR-based visual and auditory warnings with non-invasive wearable sensors to capture real-time autonomic physiological signals—particularly electrodermal responses (EDR). Contribution/Results: We present the first EDR dynamic response model for VR-AR collaborative settings, demonstrating that EDR peak amplitude and central-tendency features robustly differentiate mild from moderate occupational stress states. Machine learning identified three discriminative EDR metrics, achieving 89.2% classification accuracy. Statistical analysis confirmed that AR warnings significantly enhance EDR reactivity under moderate workload (p < 0.01). These findings establish an interpretable, physiology-grounded foundation and a technical paradigm for real-time stress monitoring and adaptive safety interventions in authentic construction environments.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Humans and AI: Brain-Sensing and Analysis
• Cognitive Modeling & Cognitive Systems: Affective Computing

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

This study examines stress levels in roadway workers utilizing AR-assisted multi-sensory warning systems under varying work intensities. A high-fidelity Virtual Reality environment was used to replicate real-world scenarios, allowing safe exploration of high-risk situations while focusing on the physiological impacts of work conditions. Wearable sensors were used to continuously and non-invasively collect physiological data, including electrodermal activity to monitor stress responses. Analysis of data from 18 participants revealed notable differences in EDR between light- and medium-intensity activities, reflecting variations in autonomic nervous system activity under stress. Also, a feature importance analysis revealed that peak and central tendency metrics of EDR were robust indicators of physiological responses, between light- and medium-intensity activities. The findings emphasize the relationship between AR-enabled warnings, work intensity, and worker stress, offering an approach to active stress monitoring and improved safety practices. By leveraging real-time physiological insights, this methodology has the potential to support better stress management and the development of more effective safety warning systems for roadway work zones. This research also provides valuable guidance for designing interventions to enhance worker safety, productivity, and well-being in high-risk settings.

Problem

• Assessing stress levels in workers using AR-assisted safety warnings
• Analyzing physiological impacts of work intensity via electrodermal activity
• Developing real-time stress monitoring for improved safety in work zones

Innovation

• AR-assisted multi-sensory warning systems for stress monitoring
• High-fidelity VR replicates real-world high-risk scenarios
• Wearable sensors collect non-invasive physiological data