@article{Zhao2026, 
author = {Long Zhao},
title = {Driver Visual Workload Characteristics and Lafety Analysis under Ramp Tunnel Conditions},
year = {2026},
journal = {Chinese Journal of Underground Space and Engineering},
volume = {22},
number = {4},
pages = {1533-1540},
keywords = {ramp tunnels, driving safety level, visual information perception, factor analysis method, SHAP analysis},
url = {https://www.sciopen.com/article/10.20174/j.JUSE.2026.04.40},
doi = {10.20174/j.JUSE.2026.04.40},
abstract = {With the extension of highways to mountainous areas, ramp tunnels have inevitably emerged. Ramp tunnels, as a special type of tunnel structure, present unique safety concerns that involve both physical factors such as linear characteristics (radius, curvature), and speed-related aspects. Additionally, visual challenges are caused by the black – white hole effect. Consequently, research on driving safety within ramp tunnels remains insufficient. In order to explore the influence of ramp tunnel alignment on driving safety, based on the driver’s information perception characteristics, key visual indicators, including fixation deviation, average fixation duration, blink frequency, saccade amplitude, and saccade speed, were collected through real-vehicle experiments. To establish a comprehensive driving safety model based on factor analysis and multiple visual indicators, and to evaluate the driving safety risks in ramp tunnels, and factor analysis method was used to quantify the driving safety, and combined with XGBOOST-SHAP, the influence of different physical quantities of ramp tunnel on ramp tunnel driving safety was revealed. The results show that: Saccade speed and vertical fixation deviation have the greatest influence on driving safety, followed by horizontal fixation deviation and average fixation time, and blink frequency and saccade amplitude have the least influence. The driving safety of the tunnel entrance section is reduced to the minimum value of 0.07 due to the ‘black hole effect’. The interaction between the linear parameters and the dynamic light environment dominates the safety variation. The curvature parameter (SHAP value 0.43) and the road illumination change rate (SHAP value 0.35) jointly explain 72% of the safety fluctuation, while the influence weight of the design speed parameter is only 11%.}
}