Incidents detection on city roads

Abstract

Abstract An important issue in traffic management is the timely and reliable incidents detection. The methods for detecting incidents are constantly being improved due to the intelligent transport systems’ development and the new sources of information emergence. Due to the incident detection tasks’ complexity, motion modeling is essential for testing various hypotheses. The article discusses the features of identifying incidents on the road network with traffic control. In the network sections, the occurrence of incidents of various durations and in different sections was simulated. The study aims to identify the minor incidents. Based on neural networks, the situations are classified and the possibilities of this approach are established. The incident detection results are presented.