Reducing violation behaviors of pedestrians considering group interests of travelers at signalized crosswalk

Violating traffic signals behaviors of Pedestrians are common phenomenon. How to reduce or even eliminate these dangerous behaviors has attracted deep attention from researchers and governments. This paper proposes signal control algorithm considering dual-flow (SCACDF), which prioritizes the interests of travelers groups and aims to use real-time traffic data to predict future short-term traffic conditions, redistribute the green time for pedestrians and vehicles, timely meet the traffic demand of pedestrians and reduce the probability of traffic accidents. A typical signalized crosswalk is employed as a case study. Experiments, which are implemented in AnyLogic software, are conducted to investigate the impact of the proposed algorithm under different levels of traffic demand (pedestrian demand and vehicle demand) scenarios. The results show that compared with traditional timing signal control, the proposed algorithm could reduce pedestrian violation rates and improve traffic safety in multiple scenarios. Under low vehicle density, the pedestrian violation rate dropped respectively by 28.8% and 45.3% on average when the green splits ratio is 0.5 and 0.7. Especially when the green splits ratio is relatively large, the pedestrian violation rate drops from 22.6% to 14.3% under the proposed control algorithm. In most cases, the re-segmentation of the red-man phase under the proposed algorithm could have a certain negative impact on vehicle traffic per unit time. However, when the green splits ratio is 0.5, the algorithm even improves the total traffic volume per unit time including pedestrians and passengers. In addition, the factors that affect the proposed algorithm and the optimal scope of application are also briefly discussed.