The Application of Wavelet Analysis to Study the Characteristics of the Traffic Flow

When introducing effective methods of traffic management using intelligent transport systems, it is necessary to take measures to obtain reliable information about the current state of traffic flows. A common problem in obtaining transport flow characteristics from technical devices is the loss of source data, which leads to the need to solve the problem of analyzing non-equidistant time series. An effective approach to the study of non-equidistant data is spectral analysis, in which the non-equidistant process is reduced to a uniform one. In this case, the restoration of missed counts leads to the appearance of a dating error. When analyzing non-stationary non-equidistant time series obtained by means of photo and video fixation, the use of wavelet analysis is proposed. Wavelet analysis is applied to transport flow characteristics that describe the dynamic state of the transport network within of a city or region. The wavelet analysis of traffic flow characteristics was carried out using Morlet wavelets on the example of three different road sections in the city of Belgorod in terms of traffic intensity and speed. The data set for analysis is a weekly interval from Monday to Sunday. The analysis of data on the average speed and number of vehicles on the section of the road network was performed. Wavelet spectra and scalograms were constructed and analyzed. Dependences of the location of extremum points are obtained, and differences are obtained during considering the spectral power.