| چکیده انگلیسی مقاله |
Accidents modeling, and, in particular, their prediction is among most important research fields in traffic engineering. By now, the relationship between traffic factors and the risk of accident occurring in urban arterials has not been fully investigated in Iran. In this regard, some works have been done in other countries such as the United States, England, Denmark, Jordan, etc., and some two-step or multi-step models and other regression models have been proposed. In this research, a statistic model is proposed to estimate probability of accident incidents in urban road segments using experiences of other countries with respect to local conditions of Iranian urban roads and through collecting traffic statistic data. To do so, factors involving in accidents in Tohid Tunnel, Tehran, Iran, and their interactions with mathematic models were investigated in the present study. Here, 4 factors in the model were considered as independent variables including: time interval, longitudinal space, following vehicle speed, and leading vehicle speed. To analyze collected statistic data, SPSS software package was applied through which logistic regression statistical and multivariate regression analyses were simultaneously applied to determine the share of each parameter in probability of accident incidents. Since validity of the collected data was approved using statistical tests and considering dynamicity of the proposed model, (evolved polynomial regression) EPR was applied as mathematical equation in this work. At the end, the final proposed model was validated using data collected from Resalat Tunnel, located in Tehran, Iran, which indicated considerable precision. |