Scientists from Perm National Research Polytechnic University (PNRPU) have developed a software package that predicts the formation of traffic jams 20–40 minutes before they appear. As reported to TASS in the press service of the university, the tool analyzes data from street photo and video recording systems, tracking the intensity of the flow in real time, and has no analogues in Russia and abroad.
The program identifies four states of traffic flow — from free movement to a formed traffic jam — and calculates the average travel time of the section. If the value exceeds the average daily value, any careless braking can trigger a chain reaction of stopping. The graph, built on the basis of 5-, 15- and 30-minute intervals, visualizes the dynamics and allows you to detect the moment when cars start to slow down more often than usual. Testing at real intersections confirmed the operability of the methodology.
The calculation results are aimed at city services and transport engineers. Knowing about the impending compaction of the flow, you can increase the green phase of the traffic light in the problem direction or bring out a traffic controller in real time. This approach replaces the reactive extinguishing of already stuck traffic with preventive management, reducing the total delay time.