Quantifying environmental modulation of arousal state in road tunnels: Implications for lighting, geometry, and
Can Qin1,2,3, Jinghang Xiao1,2,3, Longfei Cheng1,2,3
1Institute of Smart Construction and Engineering Safety, Chongqing Vocational Institute of Safety Technology, Chongqing, China.
Objective:
Road tunnels pose safety risks through abrupt luminance transitions and visual monotony that induce arousal fluctuations, yet the joint effects of lighting and geometry remain unclear. This study aims to quantify these effects via an integrated framework linking environmental conditions to drivers' arousal states.
Methods:
This study developed an integrated arousal-perception (IAP) framework to explain how tunnel environmental conditions modulate drivers' arousal states and associated safety risks. Four indicators (pupil area, heart rate, reaction time, and visual recognition distance) were collected in three urban road tunnels and integrated via principal component analysis into a composite indicator, the dynamic psychophysiological perception (DPP) model, which serves as a unified measure of arousal state. Analysis of variance (ANOVA) was used to test the sensitivity of DPP to spatial and temporal factors, Gaussian mixture modeling (GMM) to classify DPP into discrete arousal states, and structural equation modeling (SEM) to validate the causal pathways proposed in the IAP framework.
Results:
A clear spatio-temporal pattern of arousal modulation was observed. Under daytime conditions, drivers showed transient hyperarousal at threshold and exit zones, but hypoarousal in the interior under prolonged visual monotony. Nighttime conditions produced smaller oscillations but more stable activation. Differences were observed across tunnels that the selected linear-long tunnel showed the strongest arousal oscillation, whereas the spiral-short tunnel maintained more stable, elevated levels within the observed samples. The DPP model effectively captured these variations. ANOVA confirmed significant main effects of tunnel zone and time of day on DPP, while GMM classified arousal into three states (hypo-, optimal, and hyperarousal) with data-driven thresholds of 0.52 and 0.80. SEM confirmed that the DPP construct varies systematically across spatial zones and temporal conditions, supporting the IAP framework, with visual recognition distance showing the strongest loading.
Conclusion:
Road tunnels function as environmental arousal modulation systems in which lighting and geometry are associated with drivers' psychophysiological states. The DPP model provides a useful tool for quantifying arousal dynamics, and the IAP framework offers a basis for portal transition control, interior monotony mitigation, and geometry-sensitive safety optimization.