Authors: Xianfeng Yang, Yang (Carl) Lu, Gang-Len Chang
Conference: 2013 Transportation Research Board 92nd Annual Meeting
Abstract: This study presents two models for proactive VSL control on recurrently congested freeway segments. The proposed basic model uses embedded traffic flow relations to predict the evolution of congestion pattern over the projected time horizon, and computes the optimal speed limit. To contend with the difficulty in capturing driver responses to VSL control, this study also proposes an advanced model that further adopts Kalman Filter to enhance the accuracy of traffic state prediction. Both models have been investigated with different traffic conditions and different control objectives. Our extensive simulation analysis with a VISSIM simulator, calibrated with field data from our previous VSL demonstration site, has revealed the benefits of the proposed VSL control model, compared with the case without VSL. Also the results indicated that both proposed proactive models can outperform the basic models and significantly reduce the travel time as well as number of stops over the recurrent bottleneck locations. With respect to several selected MOEs, such as average number of stops and average travel time, it has been found that the one with the control objective of minimizing speed variance clearly outperforms other models.