Enhanced microsimulation framework for right-turning vehicle-pedestrian interactions at signalized intersection

Abstract

The unclear understanding of right-turning vehicle behavior at signalized intersections complicates the interaction with pedestrians. Current micro-dynamic modeling research falls short of effectively simulating this complexity. Specifically, the existing models fail to adequately capture the three states that right-turning vehicles may undergo: car-following, free right-turn, and avoidance of conflicting pedestrians. Moreover, pedestrian behavior is typically influenced by encountering conflicting vehicles and surrounding pedestrians, as well as traffic signals. To simulate these behaviors, the right-turning and yielding intelligent driver model (RTYIDM), the modified social force model (MSFM) considering green light pressure, and the yielding decision model between pedestrians and vehicles have been established. Model calibration is performed using detailed behavioral data collected and extracted from field observations. Furthermore, a microsimulation platform with 3D visualization and playback features has been developed to facilitate testing and demonstration. Model validation is performed by comparing it with actual trajectories in three representative scenarios of pedestrian crossing with conflict between pedestrians and vehicles. Meanwhile, the calibrated model's ability to predict pedestrian-interaction events and estimate vehicle yielding rates is also assessed. The well-established simulation performance of the proposed model makes it a useful tool for evaluating existing traffic operations.