Real time model-based traffic control

Subject

Traffic controllers of different control measures developed based on different type of models.

The objective of my PhD research is to reduce traffic congestion and enhance traffic operation efficiency on crowded traffic networks by developing model-based traffic controllers which can be implemented in real time. This is achieved through a number of different projects, which are summarized as:

  • A dynamic traffic assignment method that run efficiently within a few seconds for medium-sized multi-destination traffic networks. This method can be used for real time route guidance control with the help of wireless vehicle to vehicle communication and vehicle to infrastructure communication.
  • Development of a calibrated and validated macroscopic traffic flow model for freeway stretches. The model is capable to accurately predict the traffic sate in short term, say 30 minutes ahead, given the current traffic state of the freeway. The model has a better accuracy in predicting the propagation of jam waves than other models which have the same type.
  •  An optimal controller of variable speed limits to resolve freeway jam waves. Based on the current traffic state and the prediction, control signals are computed by solving a linear optimization problem. The optimization can be solved within a few seconds for a 10 km freeway stretch equipped with the variable speed limit sign every 500m.
  • Development of coordinated ramp metering approach based on the macroscopic fundamental diagram (MFD). Freeway MFD has significant hysteresis if the loading and unloading of traffic are not in the same pace. A coordinated ramp metering approach is designed to against the hysteresis.
  • Development of a hierarchical control approach to improve the traffic performance of  large-scale freeway networks in real time. The proposed coordinated ramp metering approach will be operated at the lower level, and connecting flows between freeway sub-networks are optimized at the upper level. The control approach will be tested by a synthetic large-scale freeway network to demonstrate its efficiency in computation and reducing traffic congestions.

Yu Han

Start/end date: 23-09-2013 - 23-09-2017
Daily supervisor: Yufei Yuan, Andreas Hegyi
Promotor: Serge Hoogendoorn
/* */