Concept of increasing traffic capacity by regulating traffic flows based on machine vision

Introduction. The main problem facing modern cities is the insufficient road capacity within the operating urban transport network due to the increase of traffic intensity observed over the past decades and the growing number of passenger vehicles in traffic flows. The identified problem is relevant for large and major cities across the country and it is caused by the rising level of urbanization. Regular traffic congestion worsens living conditions, negatively affecting both the economic well-being and the ecological situation in urban areas. The traffic capacity of street-and-road network depends on many factors. When addressing the issue of increasing traffic capacity, it is necessary to consider the type of transport facility — whether it is an urban road or a road of general use; if the research object is a straight section, a junction, or an intersection (built at the same or different levels). One of the tools for increasing traffic capacity today is Intelligent Transport Systems (ITS). Along with ITS, machine vision technologies have been widely used for monitoring and control tasks. The purpose of this research is to develop a concept for an intelligent transport system aimed at increasing the traffic capacity of the urban road network through regulation of traffic flows with machine vision technology. Materials and methods . A section of urban street-and-road network located in the Central District of Omsk has been investigated. One-hour surveys were carried out on weekdays, in the morning and evening peak periods. Data collection involved recording of both pedestrian and vehicle movement for a cross-section. The use of video recording made it possible to determine more accurately the number of vehicles that pass through the cross section. The data obtained formed the basis for building a transport model as a digital twin, taking design solutions, analysis of traffic census and updating traffic light cycles. The PTV Vissim software package was used to evaluate traffic capacity. The microscopic model of traffic flow includes a movement pattern for following a vehicle ahead and a pattern for changing lanes. The results of the simulation include animated traffic movements visualized online, displaying various transport and technical parameters. Results . The concept suggests redirecting traffic flows to an alternative route with machine vision, artificial intelligence and road signs of 5.15 group. Within the PTV Vissim software, a model of the proposed concept was developed. Simulation results identified the optimal operational time interval for the concept — a 20-minute period. This interval allows redirecting the flows and preventing congestions during rush hours. Discussion and conclusions . The development of intelligent systems and advanced traffic monitoring and control methods is a vital area of modern urban transport research concerning the increase of traffic capacity. For practical implementation, the proposed concept requires financial support (government funding or grants, private investments), recruiting professional experts and carrying out a more detailed project elaboration.