An analysis of road traffic flow characteristics using wavelet transform

Abstract

Measures to obtain reliable information about the current state of traffic flows are necessary to introduce effective control methods offered by modern intelligent transport systems. We developed a method and software for the wavelet analysis of road traffic flow characteristics in the frequency and time domains without restoring the missing samples. The developed method was implemented in the form of software embedded in an intelligent transport system. The method of wavelet analysis of road traffic flow characteristics takes into account the non-equidistance of data, which allows the construction of a time-frequency scan with a uniform representation without restoring the missing samples with adjustment of the sampling intervals. Background data on traffic flows for analysis was obtained from the CityPulse Dataset Collection. We analyzed such characteristics as average speed and vehicle count. We analyzed wavelet spectra and scalograms, identified common dependencies in the frequency distribution of extremes, and revealed differences in spectral power for different road segments.