Simulation of Connected Vehicle Trajectory-based Adaptive Control Algorithm in Michigan Estimated An Up to 16.3 Percent Reduction in Total Delay, Compared to Traditional Actuated Signal Control.
Trajectory Based Traffic Control Model Developed for Low Penetration of Connected Vehicles.
Ann Arbor, Michigan, United States
Trajectory Based Traffic Control With Low Penetration of Connected and Automated Vehicles
Summary Information
Connected Vehicle (CV) technologies make it possible for real-time vehicle trajectory data to be reported to traffic control systems to better understand traffic conditions. However, most of the existing CV-based traffic control models require a CV penetration rate of around 25 percent. This study developed a probabilistic delay estimation model of vehicle trajectory based real-time traffic control under a ten percent penetration rate of CVs. An adaptive signal control algorithm based on dynamic programming was implemented utilizing estimated delay to calculate the performance function representing the total delay. The proposed model was also validated through microsimulation of a real-world intersection at Huron Parkway and Plymouth Road in Ann Arbor, Michigan, considering different demand levels and CV penetration rates.
METHODOLOGY
In this study, a typical performance index of total vehicle delay was considered to evaluate the effectiveness of the proposed CV trajectory based real-time traffic control in comparison to traditional actuated signal control. The last stopped CV and the first non-stopped CV were considered to set the boundaries for the queue length approaching the intersection considered for the microsimulation model in this study. For those signal cycles that don't have any CV observed, an average hourly volume was used to generate vehicle arrival and departure times for delay estimation. Four possible cases were identified according to the existence of observed CVs; i) No Observed CV, ii) Only Stopped CV, iii) Only Non-stopped CV, and iv) Both Stopped and Non-stopped CV. The simulation study compared four CV penetration rates (0 percent, 2 percent, 5 percent, and 10 percent) with actuated signal control.
FINDINGS
- Simulation results from the real-world intersection demonstrated that the proposed model worked under 10 percent penetration rate across all scenarios. An estimated total delay reduction of up to 16.3 percent was found under congested conditions for 10 percent CV penetration rate, compared to actuated signal control.
- Under medium demand conditions, up to 5.23 percent reduction was observed in total delay, compared to actuated signal control.
- In addition, the results indicated that under congested demand level, even two percent CV penetration led to a delay reduction of 8.35 percent.
