Robust, Optimal, Predictive, and Integrated road traffic control


Our current road networks are equipped with advanced traffic control systems, such as traffic lights, ramp metering installations, and variable message signs. However, these systems are currently not exploited to their full potential. Due to this, road traffic network throughput is not optimal.

The aim of this research is the development of control strategies for variable speed limits, ramp metering installations, and traffic lights at intersections for the improvement of the throughput of regional road networks. The idea is to use traffic engineering knowledge about the traffic process and translate this knowledge into control strategies. The controller that will be developed should be able to optimize the control strategy with limited computation time and it should take the effect on the future traffic process in to account.

Scientific challenges

The main challenge is the development of controllers which are simple enough so that the control strategies can be computed with limited computation time while keeping enough complexity so that the control action still makes sense. This is not a trivial task and a lot of complexities have to be taken into account. For instance, the dynamics of traffic are usually non-linear and the description of the traffic process changes when the traffic situation changes. Additionally, the description of the traffic process on urban roads and on freeways requires different traffic flow models. Taking, among other things, these various sources of complexity into account is a big challenge.

Societal relevance

This research contributes to our understanding of the traffic process and the development of control strategies that help to make more efficient use of our infrastructure. In this way, people will experience less delay when making trips resulting in reduced travel costs and thus a societal gain.

Goof van de Weg

Start/end date: 01/10/2013 – 01/10/2017
Daily supervisor: Andreas Hegyi
Promotor: Serge Hoogendoorn