Cellular Automaton Models for Replicating Pedestrian and Motor Vehicles Interaction Behaviors at Crosswalks of Signalized Intersections.

• Main Authors: Wang, Yu-Hsing, 王佑星
• Other Authors: Chiou, Yu-Chiun
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/hseuep

碩士 === 國立交通大學 === 運輸與物流管理學系 === 107 === Conflicts and collisions between motor vehicles and pedestrians are prevailing in urban streets, especially for the pedestrian crosswalks at intersections. Without clear and accurate analysis on intertwining behaviors between vehicles and pedestrians, it is impossible to propose effective countermeasures to improve the safety and efficiency of pedestrian crosswalks at intersections. Based on this, this study aims to develop a traffic simulation model to replicate the traffic behaviors of vehicles and pedestrians at pedestrian crosswalks at signalized intersections under shared signal phases based on cellular automaton (CA) models. As to the traffic behaviors of vehicles, this study extends the refined CA model, proposed by Lan et al. (2010), which is able to accurately simulate traffic behaviors of cars and motorcycles, by adding behavioral rules for entering/leaving turning curves and car following in response to pedestrians at crosswalks. As for pedestrians, the floor field CA model, proposed by Kirchner and Schadschneider (2002), is revised for grid-based pedestrian movement at crosswalks, including walking space, step size, as well as possible moving directions. Meanwhile, the model is also modified to correctly replicate pedestrian walking behaviors attracted by certain cells and avoidance behaviors in response to obstacles at crosswalks. To integrate two parts of traffic behavior models of vehicles and pedestrians and to reflect the prevailing mixed traffic behaviors (cars and motorcycles) and “sneaking in gaps” behaviors in Taiwan, the parameters of the proposed model have been carefully adjusted and validated. The simulation results show the vehicle-pedestrian intersection behaviors in Taiwan are significantly different from those in other developed countries, because vehicles in Taiwan may not always yield to pedestrians. In order to simplify the proposed model, some unrealistic assumptions of vehicles and pedestrians can be further relaxed in future studies. For examples, this study assumes that all pedestrians are homogeneous and limited to the space of crosswalks without possibility of overflow. Additionally, only two types of vehicles are included in the CA model, cars and motorcycles. Other types of vehicles buses, trucks, and bikes can be further included.