Abstract

Research was done to examine comprehensively the safety impact of rainy weather conditions on multivehicle crash frequency and severity and to validate the impact on traffic operations through microsimulation modeling. Three primary tasks were performed to meet these objectives. For weather data processing, available data were used to estimate the following factors: rainfall intensity, water film depth, and deficiency of car-following distance. For statistical modeling, negative binomial regression was used for crash frequency, and sequential logistic regression was tested with forward and backward formats for crash severity. A better format for the crash severity estimation was determined by combining all model performance measures. VISSIM was used to design traffic simulation models to reflect the effect of weather on traffic operation with five scenarios of the following weather-sensitive parameter adjustments: desired deceleration rate function, desired speed distribution, and headway time. As weather-related determinants, daily rainfall and wind speed were found to be statistically significant to crash frequency and severity estimations, respectively. VISSIM provided the most similar traffic data to the observed data when both desired speed distribution and deceleration rate function were adjusted. Statistical modeling in this research can be used to examine highway safety in rainy weather and to provide quantitative support on implementing road weather safety management strategies. Correspondingly, the adjustments of weather-sensitive traffic parameters will be the preliminary step to measure the strategy efficiencies through safety surrogate indexes in traffic simulation.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call