Прогнозування ризиків ДТП на автомобільних дорогах

Abstract

The issue of road safety is particularly relevant in Ukraine because according to official statistics, every 18 weeks accidents occur on the country’s roads, and every 117 minutes a person is killed by an accident. Road traffic crashes are random events that may be subject to a certain pattern. Therefore the analysis of crash patterns, their distribution in time and space, and the identification of causes is an urgent issue that has a significant impact on the entire transport system of the country. In the presented work we considered the assumption that the distribution of crash rates is subject to a certain distribution law. Checking the compliance of a random variable with one of the distribution laws was carried out for the condition that the level of road safety expressed by the final coefficient of accidents. The use of this indicator is conditional on a wide range of traffic condition parameters being taken into account, making it suitable for use in road safety audits. The study was carried out using databases of road safety values on public roads, which are available in the Road Traffic Management and Safety Department and are periodically updated. As a result of the study of the distribution of traffic accident risks, the coincidence of the nature of the theoretical distribution of the considered values of the real (relative accident rate) and predicted levels of road safety (the final accident rate (express model)) was found. According to the obtained analytical dependences of the mathematical expectation of the occurrence of a road accident we can determine the center of accidents with a certain probability, and the variance of the lognormal distribution describes the dispersion of a random variable (the number of road traffic crashes) on a selected road section. The resulting models offer the possibility of using them to determine road crash risks on difficult and dangerous sections of public roads.