Assessing Workers Neuro-physiological Stress Responses to Augmented Reality Safety Warnings in Immersive Virtual Roadway Work Zones

TL;DR

Multisensory AR warnings in immersive VR roadway work zones elicit distinct neuro-physiological responses captured via $HR$, $HRV$, $EDA$, and $EEG$. The study uses a within-subject design with light and moderate tasks and analyzes the timing and magnitude of neural (alpha/beta) and autonomic responses to warnings, identifying key markers such as $Mean\text{-}EDR$ and $Min\text{-}NNI$. It finds a time lag of approximately $4.9$–$5.8$ seconds between neural changes and subsequent autonomic responses, and demonstrates that task intensity shapes brain–body coupling, enabling adaptive AR safety interventions and real-time stress monitoring. These findings support designing individualized AR safety systems that balance safety benefits with cognitive load in high-risk construction environments.

Abstract

This paper presents a multi-stage experimental framework that integrates immersive Virtual Reality (VR) simulations, wearable sensors, and advanced signal processing to investigate construction workers neuro-physiological stress responses to multi-sensory AR-enabled warnings. Participants performed light- and moderate-intensity roadway maintenance tasks within a high-fidelity VR roadway work zone, while key stress markers of electrodermal activity (EDA), heart rate variability (HRV), and electroencephalography (EEG) were continuously measured. Statistical analyses revealed that task intensity significantly influenced physiological and neurological stress indicators. Moderate-intensity tasks elicited greater autonomic arousal, evidenced by elevated heart rate measures (mean-HR, std-HR, max-HR) and stronger electrodermal responses, while EEG data indicated distinct stress-related alpha suppression and beta enhancement. Feature-importance analysis further identified mean EDR and short-term HR metrics as discriminative for classifying task intensity. Correlation results highlighted a temporal lag between immediate neural changes and subsequent physiological stress reactions, emphasizing the interplay between cognition and autonomic regulation during hazardous tasks.

Figures (8)

• Figure 1: Overview of the primary methodological components: (a) Data acquisition, (b) Data analysis, (c) Inferring neuro-physiological stress metrics
• Figure 2: Experimental setup of participants performing light and moderate intensity tasks in the immersive virtual reality environment of roadway work zones, and multi modal safety warning delivery mechanisms
• Figure 3: (a) Utilized EEG device, spatial configuration of EEG electrodes, and targeted brain areas (b) Utilized wearable wristband and its embedded sensors
• Figure 4: Comparative analysis of post-warning HR and HRV features between light and moderate activities
• Figure 5: Comparative analysis of post-warning EDR for light and moderate activities