Exploring the Mechanism for Increased Risk in Freeway Tunnel Approach Zones: A Perspective on Temporal-spatial Evolution of Driving Predictions, Tasks, and Behaviors

Abstract

Freeway tunnel approach zones, situated outside the tunnel, do not undergo the same sudden changes in luminous environment and visual references that entrance zones experience. Despite this, accident data indicates that approach zones present similar safety risks to entrance zones, both of which are significantly higher than other tunnel sections. The reasons for the heightened risks in approach zones remain unclear in existing research. To address this knowledge gap, this study conducted real vehicle tests and subjective perception experiments. The Task Analysis of Driving Scenarios (TADS) was employed to analyze driving behavior patterns and develop a set of evaluation metrics, including four key driving behavior nodes (P1_SGD, P2_EF, P3_FF, P4_SAD), safety and efficacy indices for active deceleration behaviors (I1_ADS, I2_ADE), and two indicators for understanding anomalous behaviors (SR, AOI_PFN). By skillfully selecting scenarios to control variables, this research examined how limited visibility in tunnel approach zones and spatial intervisibility tunnels contribute to safety risks in these zones. Additionally, the Predictive Processing Model (PPM) was used to elucidate the temporal and spatial evolution of driving predictions, tasks, and behaviors under normal conditions. The findings reveal that, although heavy driving tasks cannot be avoided, under normal conditions, predictions develop gradually with minimal prediction errors, enabling effective navigation. However, limited visibility in tunnel approach zones and spatially intervisible tunnels lead to inaccuracies and deviations in predictions, resulting in significant prediction errors as drivers approach the tunnel. This causes more aggressive driving behaviors, disrupting the delicate balance of predictions and tasks.