Nighttime safety evaluation for signalized intersections at the signal cycle level based on rear-end conflict models considering lighting and traffic conditions

Objective: The primary objective of this study is to explore the effects of the lighting level on nighttime safety of signalized intersections based on conflict models under traffic conditions varying in cycles. Method: Nighttime data were collected from a field study at six signalized intersections in Nanjing, Jiangsu Province in China. Nighttime rear-end conflict models were developed by adopting the generalized linear model (GLM) approach to relate the frequency of rear-end traffic conflicts to lighting level, traffic volume and platoon ratio at the signal cycle level. Results: The final model consisting of all explanatory variables, including lighting level, traffic volume, and platoon ratio demonstrates a better performance of safety evaluation when compared to the model considering traffic volume only and the model with traffic volume and an additional variable of lighting or traffic conditions. Nighttime safety of signalized intersections is expected to improve with larger platoon ratios and higher lighting levels. A potential application of the final model was further explored by benefit-cost analyses. The analyses provided a hypothetical recommended lighting level under various traffic volumes and platoon ratios when safety benefit equals lighting system cost. Conclusions: Nighttime safety can be evaluated using the developed rear-end conflict models, which relate the number of rear-end conflicts to traffic and lighting variables. The number of rear-end conflicts can be calculated by the final conflict model with lighting level, traffic volume, and platoon ratio. The developed model can be potentially applied to provide further insights on the lighting management for intersection safety optimization with traffic conditions varying in signal cycles via vehicle-to-infrastructure (V2I) communications.