عنوان مقاله [English]
Walking is a prominent place in the urban transport system, as a significant part of intra-city trips takes place on foot. Increasing motor vehicles, resulting in traffic congestion, air pollution, rising fuel prices, as well as the increasing attention of people to their health, the ways of transport and motor vehicles, especially walking, have become more popular than ever before. On the other hand, due to these issues, the lack of attention to the pedestrian safety category has led to numerous road accidents over the past years. Safety analysis of paths is commonly used, often with accident statistics data, and the use of statistical models to determine accidental and insecure points. Several reasons indicate that this method is not a suitable tool for evaluating safety. Therefore, the use of alternative methods or complementary methods for non-accidental data to improve traffic safety has been considered by researchers in traffic science. One of the most important methods is the use of traffic interaction indicators. . In the present study, we tried to assess the safety of pedestrians in uncontrolled intersections in urban areas by using neural network method and alternative method based on non-random data. At first, by using the neural network method, factors affecting the occurrence of an interaction between cars and pedestrians were determined and finally, by using Post Encroachment Time (PET) and Time-To-Collision (TTC) indicators of critical thresholds the probability of automobile-pedestrian interactions in uncontrolled intersections of the city was determined.