A New Dilemma Zone Protection Algorithm based on Real-Time Vehicle Trajectory

摘要

Dilemma zone (DZ) has been recognized as a major potential causing rear-end and right-angle crashes, especially at high-speed signalized intersections. The conventional DZ protection systems use advance detectors to extend the green until either there are no vehicles in the DZ or the green time reaches the maximum. Recent advances in sensor technologies and intelligent transportation systems allow us to obtain more detailed vehicle trajectory data from the field. Hence, the objective of this study is to develop a Real-Time Vehicle Trajectory based Dilemma Zone Protection Algorithm, i.e., RVT-DZ algorithm. In the proposed RVT-DZ algorithm, both the current detected number of vehicles and the predicted number of vehicles in Type I DZ and option zone are taken into account with different priorities when deciding the time of ending a green phase. Meanwhile, trucks are given prior and separate protection. To evaluate its validity, this study selected a high-speed rural intersection located in Shanghai as the study site and developed an online simulation tool to compare the proposed RVT-DZ algorithm with the conventional DZ protection algorithm, i.e., Green Extension System (GES). Results support that the RVT-DZ algorithm is capable of reducing the number of vehicles caught in Type 1 DZ more effectively, while its prevention of vehicles from being caught in Type 2 DZ is less efficient as compared with the GES algorithm. Both of the algorithms showed good performance in reducing the number of vehicles in theoption zone. In addition, the proposed RVT-DZ algorithm could improve the control delay approximately by 12% when comparing with the GES algorithm.