Analysis of the causes of pedestrian-vehicle traffic accidents based on Bayesian networks

Abstract

With the continuous improvement of people's living standards, the number of cars has also increased dramatically. While cars are convenient for people to travel, they also lead to increasingly serious traffic safety problems. For this reason, this paper uses the fault tree and Bayesian network methods to conduct an in-depth study on the causes of pedestrian-vehicle traffic accidents from three aspects: people, vehicle, road and the environment. In this paper, the occurrence of pedestrian-vehicle traffic accidents is divided into 29 basic events. The basic events of each of the 381 pedestrian-vehicle traffic accidents were Classified by 0–1. A fault tree model leading to pedestrian-vehicle traffic accidents is established, which is then transformed into a Bayesian network model, and Bayesian network inference, sensitivity analysis is performed with the help of Netica software. Our results suggest that illegal crossing of traffic lanes, speeding, rainy day, slippery road, braking is not timely, visual impairment are the main causes of pedestrian-vehicle traffic accidents. These results can not only provide a reference for transportation technology, but also provide a basis for government legislation.