Coordinated Arterial Dilemma Zone Protection Through Dynamic Signal Timing Optimization

Vehicles trapped in dilemma zones on arterial streets are at high risk of rear-end collision or right-angle collision. Most of extant studies focus on dilemma zone protection at isolated signalized intersections. A coordinated arterial dilemma zone protection model at the lane group level, aiming to minimize the weighted combination of safety and efficiency performance indicators, is proposed to improve the safety of arterial traffic by dynamically optimizing the timing parameters of arterial signals. The problem is solved with the rolling optimization technique which enumerates and evaluates feasible green phase durations for all lane groups over each time horizon by embedding the cell transmission model for traffic prediction and the genetic algorithm for signal timing optimization. The effectiveness of the dynamic optimization method is verified by comparing it with actuated control and isolated and coordinated fixed-time control strategies. The results show that the proposed method is able to obtain better safety performance by tending to choose longer cycle lengths and that the proposed method is more suitable for situations where traffic demand changes drastically. With the aid of a traffic surveillance system, the proposed model can be used to provide coordinated dynamic dilemma zone protection for arterial traffic.