Abstract:

In order to accurately identify vehicle interaction risk in urban tunnel merging zones and to reveal the dynamic evolution of risk driven by the interaction among tunnel environment, vehicle behavior, and driver characteristics, a comprehensive risk-field model was proposed. The model incorporates tunnel environmental constraints, vehicle risk propagation effects, and driver risk perception characteristics. An empirical study was conducted using a typical tunnel merging zone in Chongqing as the case study. First, two typical acceleration-lane configurations, namely direct and parallel types, were examined. Each merging zone was divided into four critical sections: the entering, converging, merging, and mainline sections. Second, based on driving data collected from real-world vehicle tests, the distributions of vehicle speed across different sections, driving behavior patterns, and variations in drivers' risk perception were systematically analyzed. The mechanisms through which these factors contributed to driving risk formation were then examined. Finally, the spatiotemporal distribution of driving risk in different tunnel merging zones was compared and evaluated. Risk visualization was also performed, and differentiated risk prevention and control strategies were proposed. The results show that driver risk perception reaches its highest level in the merging section and varied significantly across sections. Driving risk in both types of tunnel merging zones is highest in the merging section, followed by the mainline section, and then gradually decrease in the converging and entering sections. Compared with the direct acceleration lane, the parallel acceleration lane shows reductions in risk-field strength of 8.22%, 8.71%, 9.35%, and 6.57% in the entering, converging, merging, and mainline sections, respectively, indicating a lower overall driving risk level.

Key words: urban tunnel confluence zone, risk perceptionl, driver, perception characteristics, comprehensive risk field