Safety evaluation of visual load at entrance and exit of extra-long expressway tunnel based on optimized support vector regression

Abstract

The traffic environment of an extra-long expressway tunnel is more complex than that of a long tunnel, which increases the driving risk. The visual load of drivers can be used to evaluate driving safety and comfort. To reveal drivers’ visual load characteristics at the entrance and exit of extra-long tunnels on mountainous expressways, this study conducted vehicle tests with 12 drivers at Gonghe extra-long tunnel on the Yu-Xiang expressway in the Wulong District. An eye tracker, non-contact multifunctional velocimetry, illuminometer, and other test equipment were used to record drivers’ pupil areas, velocity, and illuminance when entering and leaving the tunnel. The change characteristics of drivers’ pupil areas were studied. The maximum transient velocity value (MTPA) of the pupil area was selected as an index to evaluate the visual load degree. Based on velocity and illuminance coupling, a visual load model was constructed using the optimized support vector machine (GA-SVM). The influence of velocity and illuminance on the MTPA in the tunnel’s approach, entrance, exit, and departure section was analyzed. The results show that drivers’ psychological tension order at the entrance and exit is entrance section ≈ exit section > departure section > approach section. In the approach section, the visual load is mainly affected by environmental illumination. In the entrance and exit sections, the visual load is positively correlated with velocity and negatively correlated with illuminance, and velocity has a greater impact on visual load. In the tunnel departure section, the two variables synergistically influence the driving visual load. The research results provide theoretical support for the safety design and management of extra-long tunnel entrances and exits.