Adaptiv signalstyring i realtid

Studenteropgave: Speciale

  • Mikkel Færgemand Hansen
4. semester, Veje og Trafik, Kandidat (Kandidatuddannelse)
In recent years, the traffic in the cities has increased significantly because of urbani-zation and a rising number of vehicles per inhabitant. The increasing traffic has caused congestion in the cities, which reduces the mobility and accessibility. This has led to an increased focus from the politicians, and the aim is to make the existing infrastructure more efficient. In this connection, the focus is at the signalized intersections because they set the limit of the capacity.
The aim of this project is to create a new and intelligent signal controller, which can execute the traffic efficiently compared to traditional signal control. First of all the tra-ditional principles such as maximum green time, maximum cycle time, predefined order of signal phases and offsets between coordinated signalized intersection is replaced with one principle where the overall traffic situation is attempted opti-mized.

Despite the principles, a new detection technology is used in the project. Instead of detection in some given points towards the intersection, the new technology allows for a continuous object detection. With this technology, it is possible to calculate the ETA (Estimated Time of Arrival), which is an essential parameter in the intelligent signal control.

The data from the continuous object de-tection is used in a mathematical tool, Uppaal Stratego. Uppaal Stratego use machine learning and the purpose is to optimize the overall traffic around the sig-nalized intersections.

The project location is a major road in Aalborg, Denmark, called Hobrovej with six signalized intersections. The project contains a comparison of scenarios with the existing signal control and the intelli-gent signal control by using microscopic simulations in VISSIM. The scenarios are compared based on average delay, queue lengths, stops, fuel consumption and the overall travel time at the road.

The results based on the microscopic simu-lations show, that the intelligent controller provide a much more efficient execution of the traffic in the signalized intersections. Thus the average delay and the queue length is reduced by 27-54 % and 42-64 %, respectively. Furthermore, the intelligent signal controller reduce the fuel consump-tion and the number of stops in the signal-ized intersection. Additionally the overall travel time at Hobrovej is reduced.
Udgivelsesdato8 jun. 2017
Antal sider94
ID: 259340563