Selecting the direction of improving the traffic light system of urban traffic flows management

The number of vehicles has been increasing annually, the number of vehicles increases, the intensity and density of traffic flows increase as well, hence, the rate of road traffic injuries becomes important. If growth of an injury rate is connected, generally with non-compliance with the high-speed mode, then environmental pollution happens because of increased time of vehicles in traffic jams. Therefore, it is necessary to provide uniform promotion of traffic flows in the cities for decreasing the level of road traffic injuries and emissions in the environment. One of effective modern methods of ensuring uniform motion of traffic flow is the application of systems of a transport telematics, in particular, the management of systems of traffic signs, road boards and the traffic light alarm system. The analysis of the existing systems and methods of traffic light regulation is provided in the article. All analysed systems and methods are based on application of uniform data – data on standard parameters of traffic flows. Need of collection and the analysis of additional semistructured data on the factors exerting significant impact on parameters of traffic flows in the cities is shown in paper. As instruments of collection and the analysis of diverse data, it is offered to use the Big Data tools. The algorithm of forecasting parameters of traffic flows based on the original idea of resource flows and a combination of the Big Data methods is proposed ("the closest neighbor" and Kallman's filter) with an optimization method for searching the minimum covering tree on the resource networks describing functional dependences between the actual and forecast values of data.