The study of physical characteristics of highways according to road patterns to create accident prediction models

Abstract

The objective of this study To study the physical factors of the highway according to the road pattern That affects road accidents And create a model to predict highway accidents If the type of physical characteristics of the highway is chosen according to different road formats By dividing the physical characteristics of the highway according to the road format into 4 types. And hypothesize variables that may affect the number of road accidents according to the physical characteristics of the road that may result in accidents 36 factors. From the study, it is found that, type 1, highways, medium islands. There are 18 factors that affect the accident. The second form is the type of highway with a middle island type with a grooved type. There are 18 factors that affect the accident. The 3^rd type is highway type, middle island type, color island type. There are 6 factors affecting the accident. And the 4th pattern Characteristics of highways in the form of railways or barricades. There are 5 factors that affect the accident causing an accident on the highway. And results of variable analysis. Which has the model that has the most suitable variable to create a model to predict the occurrence of highway accidents. It was found that, model 1, the model for predicting highway accident is Number of accidents (Y) = - 6.273 + 0.317 Proportion of heavy trucks - 1.062 Number of intersections in the surveyed road range +1.627 Land use for government buildings and institutions - 0.026 Motorcycle proportion + 0.458 Slope of surveyed road + 0.508 Width of central island. The second model, the model for predicting highway accidents, is Number of accidents (Y) = 2.168 + 0.621 The width of the central island - 0.063 Proportion of motor vehicles exceeding the speed limit - 0.054 Proportion of motorcycles. The third form, the model for predicting highway accident is Number of accidents (Y) = 0.385 +1.264 Uses or unprofitable areas of transportation and transportation stations + 0.507 is characteristic of urban areas + 0.104 Central island width - 0.12 Proportion of heavy trucks. And the 4th pattern Yes, the model for predicting highway accidents is Number of accidents (Y) = 0.773 + 2.623 the distance between the right lane and the central island. Which is the most suitable model for this study.