چکیده:
Background and Aim: Traffic accident prediction models can help to understand the causes of accidents and predict their occurrence in specific situations. The aim of this study was to create a crash prediction model for Yazd-Anar highway as part of the north-south road transport corridor using fuzzy logic method and to further analyze the sensitivity of different crash reduction strategies to help make a decision. Recipients are provided to implement appropriate corrections. Method: The method presented in this research is the use of fuzzy logic to predict and analyze the sensitivity of factors affecting accidents that may occur during the year. The data of this research are Yazd-Anar highway accidents in the last ten years and the research variables include road width, road pavement conditions, average hourly traffic volume, speed, number of access points to the highway and the status of traffic signs. By performing sensitivity analysis, appropriate steps and precautions have been suggested in order to reduce accidents. Findings: The research findings showed that the proposed fuzzy logic system provides an accurate and stable prediction of the number of traffic accidents with a coefficient of determination of more than 87%. Results: Implementation of short-term and long-term programs resulting from the sensitivity analysis of factors affecting traffic accidents will reduce accidents by 15.2% and 42.3%, respectively.
خلاصه ماشینی:
This study was conducted with the aim of creating an accident prediction model for the Yazd-Anar highway as part of the north-south road transport corridor of the country using the fuzzy logic method, and subsequently, a sensitivity analysis of different accident reduction strategies has been provided to assist decision-makers in implementing appropriate reforms.
In this study, a fuzzy model for predicting driving accidents on the Yazd to Anar highway in Kerman province—which, as part of the north-south road transport corridor of the country, plays a vital role in transporting goods and cargo from southern ports to other cities of the country and vice versa, and where the excessive frequency of accidents resulting in death and injury on the mentioned highway is a fundamental problem that threatens the life, safety, and property of people—is presented as a case study.
Based on this, the average annual hourly frequency per lane 1 (AHTL) as the number of passing vehicles per hour in each lane, road width (rw), speed (sp), number of partial access points (ma) along one unit of road, road surface condition (pm), and the percentage of guidance and driving signs per kilometer (sj) were considered as the model input variables, and annual all accidents 2 (AAA) as the model output variable (dependent variable).
Kimensch and Rahmanian 1- Meng, Zheng, Qin 2- Du, Wang, Guo, Li 3- Mamdani, Assilian 4- Ghanbari, Mehr, Nehzat 5- Saravanan, Sabari, Geetha (1400: 10) also used artificial neural networks for accident prediction and finally compared the results obtained from the research using the multivariate analysis method.
(
, 
, 
)
