Features of Using the Method of Normalised Range and Fractal Analysis in Studying the Car Traffic Flow Intensity

The relevance of studying the methods of studying traffic flows is due to the need to analyse their features, determine the allowable areas of their application in solving practical problems of the transport sector. The objective of the work is to identify the results of application of modern methods of time series analysis that use the values of intensities of car traffic flows on the urban street-and-road network. The subject of the study is associated with the calculated Hurst and fractal dimension indices (fractal analysis), as well as with checking the validity of the quantitative relationship of these indicators, specified by B.Mandelbrot and used in applied research, on real data on intensity of car traffic flows. Digital data for the study were obtained using «Azimuth» stationary measuring software and hardware photo and video recording complexes, located on the street-and-road network of the city. The study has found that anomalous values of key indicators are encountered when using the normalised range method and fractal analysis: the Hurst exponent takes values outside the usually defined range, the relationship between the fractal dimension and the Hurst exponent does not fully correspond to the known B. Mandelbrot set. It seems necessary to conduct a deep and thorough study of the results obtained when using the above and, possibly, other methods for studying the intensity of traffic flows on street-and-road networks.