PNIPU Transport Flow Forecasting Methodology

Main article: Data mining Data mining

2022: Introduction of Transport Flow Prediction Methodology

A study by scientists from the Perm Polytechnic will help predict traffic flows. The technique will make it possible to more effectively manage the movement of cars in cities and reduce the number of congestion and road accidents. The university announced this on July 14, 2022. It will be possible to regulate the transport network based on the analysis of information from video recording complexes for traffic violations. According to scientists, the development will ensure the technological sovereignty of Russia.

Specialists from the Perm Directorate of Road Traffic and Road Safety Technologies LLC also took part in the development.

Using the normalized scope method, or Hirst's measure-finding, random data is studied in various fields, scientists say. For example, it allows you to predict the value of financial assets, shares and exchange quotes, market prices, currency fluctuations and economic risks. In addition, this approach is applied in physiology or medicine, in meteorology, agriculture and industrial production, as well as for the diagnosis of dynamic hydrological systems and earthquake forecasting.

We were the first Russia to use data from software and hardware complexes that record traffic violations to study traffic flow and analyze its intensity using this method. To do this, we investigated a section of the highway in Perm with intensive one-way traffic. It is equipped with video recording equipment that allows you to measure the characteristics of the flow of transport. We managed to find out how the intensity of cars changes during the week, and determine the patterns of changes in traffic flows, "said project Mikhail Boyarshinov, head, professor of the Department of Cars and Technological Machines of the Perm Polytechnic, Doctor of Technical Sciences.

Scientists have identified trends in changes in traffic flow intensity and other characteristics. The Hirst index does not evolve equally over the course of different days, but by the end of each day it takes on close values. This shows the stability of the process during the day and the presence of a trend in traffic, the developers say. At the same time, calculations show that throughout the week, the intensity of the transport flow exhibits properties characteristic of a random process.

According to scientists, the technique will help predict the movement of car flows, control the operation of traffic lights to prevent traffic congestion, and monitor equipment. This approach can be used in advanced planning, in the reconstruction of the road network and the design of new highways.