Modeling the traffic capacity of the nodes of urban transport network based on the fuzzy logic methods

Traffic state prediction is a key component of intelligent transport systems (ITS) which has attracted a lot of attention over the past few decades. The improvement in the accuracy of modeling and predicting the traffic capacity of intersections, depending on such uncertain factors as the intensity of pedestrian flow and its discontinuity, is possible only with the development and use of new methods. In order to form a number of typical control algorithms for each regulated node of a city transport network, the need to cluster them arises. The traffic flow parameters of each separate regulated node of transport network have been measured using convolutional neural networks (YOLOv3). As a result of the analysis of the differences between the clusters in terms of mean values of independent factors, statistically significant differences have been revealed and linear regression models have been detected. On the basis of these models, typical management decisions on increasing the traffic capacity of regulated nodes of the transport network will be formed. When constructing the model, the fuzzy logic methods, as they more fully reflect the influence of random factors of pedestrian flow on the traffic capacity of the intersection as a whole, have been used.