Modeling Cooperative Lane Changing and Forced Merging Behavior
Modeling Cooperative Lane Changing and Forced Merging Behavior
复制标题
DOI：
--
复制
发表时间：
2007
期刊：
影响因子：
--
通讯作者：
C. Choudhury;M. Ben-Akiva;T. Toledo;Gunwoo Lee;Anita Rao
中科院分区：
文献类型：
--
作者：
C. Choudhury;M. Ben-Akiva;T. Toledo;Gunwoo Lee;Anita Rao
合并地点是高速公路瓶颈的主要来源。微观仿真工具在分析这些瓶颈和设计最佳几何配置和控制策略方面越来越受欢迎。现有的方法没有明确考虑驾驶员之间的合作以及不耐烦的驾驶员的激进并道，并且经常过度预测拥堵。我们提出了一种组合合并模型，该模型将正常、协作和强制合并组件集成在单个框架中。该模型的参数已根据从加利福尼亚州 80 号州际公路收集的详细车辆轨迹数据进行了估计。两个更简单的模型：单级间隙接受模型和 b．还提出了显式强制合并模型，以证明在决策框架中明确包含礼貌和强制合并的重要性。估计结果的统计比较表明，与简单模型相比，组合模型具有明显更好的拟合优度。所有三个模型均在微观交通模拟工具 MITSIMLab 中实现，并根据从加利福尼亚州 US 101 收集的聚合轨迹数据进行验证，该数据具有与估计数据收集站点不同的几何配置。模型验证结果也支持行为假设，并证实组合模型的性能明显优于更简单的模型。
Merging locations are major sources of freeway bottlenecks. Microscopic simulation tools are receiving increased popularity in analyzing these bottlenecks and designing optimum geometric configurations and control strategies. Existing approaches do not explicitly consider cooperation among drivers and aggressive merges by impatient drivers and often over-predict congestion. We present a combined merging model that has normal, cooperative and forced merging components integrated in a single framework. Parameters of the model have been estimated with detailed vehicle trajectory data collected from Interstate-80, California. Two simpler models: a. Single Level Gap Acceptance Model and b. Explicit Forced Merging Model are also presented to demonstrate the importance of explicitly including courtesy and forced merging in the decision framework. Statistical comparisons of estimation results indicate that the combined model has significantly better goodness-of-fit compared to the simpler models. All three models were implemented within the microscopic traffic simulation tool MITSIMLab and validated against aggregated trajectory data collected from US 101, California, which has a different geometric configuration than the estimation data collection site. The model validation results also support the behavioral hypotheses and confirm that the combined model performs significantly better compared to the simpler models.