Intelligent Management of Traffic Flows in Large Cities

Abstract

The study is devoted to solving the problem of traffic jams that arise on the roads of large cities due to various random factors. An overview of the modern systems of automatic control of traffic flows is provided, and the main reasons for the occurrence of traffic anomalies at intersections controlled by traffic lights are considered. The formalisation of input and output data corresponding to stochastic traffic conditions is shown. The research focuses on improving the city’s traffic flow management systems by implementing distributed data processing systems. The architecture of the Distributed Data Processing System is proposed, and the scheme of its integration with the Intellectual Traffic Light Control System, which is being developed for automatic situational adjustment of traffic light operation, is shown. The practical significance of such integration lies in the ability to operate with up-to-date information about the situation on the city’s roads, which is necessary for the automatic coordination of the city’s traffic lights. This possibility can be realised thanks to the ability of distributed data processing systems to process large volumes of stochastic information in real time. In the future, the accumulation of statistical data on the state of traffic will provide an opportunity to form reliable samples for training artificial neural networks capable of processing data from video surveillance cameras.