MSc Thesis Proposal

Modeling and control of highway traffic

Mentor: B. De Schutter

Keywords: traffic, modeling, control, optimization

Traffic jams are becoming more and more acute every day. They do not only cause considerable costs due to unproductive time losses; they also augment the possibility of accidents and have a negative impact on the environment (air pollution, increased fuel consumption) and on the quality of life (noise, stress). Since traffic congestion is such a pressing problem, there certainly is a need for measures that can be implemented on the short term. One of such short term measures is to augment the capacity of the existing infrastructure by regulating and redirecting the traffic flow.

In this proposal we consider the modeling and control of traffic flows on highways. We shall start by studying several models that describe the traffic flows on the highway and then select one or more that best suit our aims. Next, we shall investigate how control measures such as dynamic route information panels, variable message signs, ramp metering (see figure below), advisory speeds, etc. can be used optimally to decrease the lengths and the frequency of occurrence of traffic jams.

Ramp metering
Ramp metering.

The operation of many contemporary traffic controllers is based on the local traffic situation. E.g., ramp metering can be switched on when the traffic around the on-ramp is becoming too dense, speed limits are imposed when there is congestion downstream, or route information is given based on congestion in the vicinity of the DRIPs (Dynamic Route Information Panels). While these control measures are aimed at solving or alleviating the local problems, they also have an effect further away (and thus later) in the network. E.g., solving a local congestion can create another congestion somewhere else in the network. Therefore, considering the future network effects of these control measures can considerably improve the flows on the network. Model predictive control seems to be a good approach to tackle this problem, since predictions about the future behavior and the development of traffic flows can be considered.

If you are interested in selecting this project as your MSc project, please come along or send an email for more information.

This page is maintained by Bart De Schutter.