Flexible lane configuration for intersections
|< Prev||Next >|
The application of flexible lane configuration for signalised intersections is sometimes beneficial for traffic delays, but sometimes not. This is the conclusion of the research conducted by Frank Anema and supported by TrafficQuest. He did his master thesis on this topic and received his MSc degree with the report entitled “Model-predictive control of the lane configuration at signalized intersections”.
Flexible lane configurationAt this moment there are five intersections in the Netherlands with a flexible lane configuration. This means that during a certain period of the day an extra lane is available for a busy movement on an intersection and the same lane is available for another movement during the rest of the day.
This is especially useful if there are busy, clear distinct movements for the same arm of an intersection during different periods of the day. For these intersections lanes are switched in practise twice a day at fixed times. Making this switching of lanes more flexible and dependent on the circumstances could lead to less delay for the intersection as a whole. This research investigates the possibilities and related effects for this.
Controllers for flexible lane configurations
During the research two controllers were developed. Both controllers use the current situation and measurements to determine whether or not lanes should be switched between movements. The first controller predicts the total delay for the intersection to make a decision, based on the known traffic demand and the different lane configurations. The other controller measures queue length to determine the switching policy.
Testing with simulation
Both controllers were tested with a simulation model for two real-life intersections. For a first scenario a fictitious demand profile for a short period was used and for a second scenario a realistic, measured demand profile for a whole day. For the first scenario the continuous checking if the lane configuration should be changed or not, was beneficial for the total delay on the intersection. This was not the case for the real-life demand profile. The situation in that case stayed the same. Further research should reveal what causes that. The report Model-predictive control of the lane configuration at signalized intersections provides several recommendations to start this research.