考虑宏-微观跨层效应的车辆换道意图影响因素分析  

Analysis of influencing factors of vehicle lane changing intention considering macro-micro cross-layer effect

作　　者：龙雪琴[1] 林海斌 杨子江 LONG Xue-qin;LIN Hai-bin;YANG Zi-jiang(College of Transportation Engineering,Chang'an University,Xi'an 710064,Shaanxi,China)

机构地区：[1]长安大学运输工程学院,陕西西安710064

出　　处：《长安大学学报（自然科学版）》2024年第6期114-123,共10页Journal of Chang’an University(Natural Science Edition)

基　　金：国家自然科学基金项目(52002033);陕西省自然科学基础研究计划项目(2024JC-YBMS-338)。

摘　　要：为实现对高速公路车辆换道意图的准确分析,揭示交通流状态和车辆运行状态双层因素对换道意图的交互作用,考虑车辆在换道过程中宏观因素和微观状态的跨层交叉效应,构建基于多层线性模型的车辆换道意图模型。以High D数据集作为样本库,根据车辆换道的横纵向运行规律提取换道和车道保持样本;结合相关性及多重共线性分析结果,选择4个对驾驶人换道意图有显著影响的车辆微观运行状态因素和2个宏观层面的交通流运行状态因素,作为多层线性模型的解释变量,分别建立零模型、随机系数模型和完整模型;使用重要性能指标对模型进行比较,验证宏-微观多水平因素对换道意图的影响。研究结果表明:驾驶人换道意图同时受到微观车辆个体层因素和宏观交通流层因素的影响;微观上,驾驶人换道行为多是由当前车道周边车辆个体的驾驶行为所引起;一定的微观驾驶行为特征下,车道平均速度增加会降低驾驶人换道意图,而车道密度增加将进一步刺激驾驶人的换道行为,微观车辆个体与宏观交通流指标之间存在着显著的跨层交互效应;考虑宏-微观影响因素的多层线性模型比单水平模型和仅单层的随机效用理论模型具有更高的拟合优度和模型精度,较真实地反映了车辆换道的交互行为,能够更好地解释车辆个体与交通流间的嵌套关系。To accurately analyze the lane changing intention of vehicles on expressways and reveal the interaction between traffic flow state and vehicle running state,a vehicle lane changing intention model based on multi-layer linear model was constructed,considering the cross-layer crossover effect of macro factors and micro states of vehicles during lane changing.With High D data set as the sample library,lane change samples were extracted according to the transverse and longitudinal running regular of vehicle lane changing.Combined with the correlation and multicollinearity analysis results,four kinds of vehicles'micro running states and two kinds of traffic flow macro running states that have significant influence on drivers'lane changing intention were selected as explanatory variables for the multi-layer linear model.The zero model,random coefficient model and complete model were established respectively,the influence of macro-micro multilevel factors on lane changing intention was verified by comparing the models with important performance indexes.The results show that the vehicles'lane changing intention is influenced both by micro factors of vehicles'individual layer and macro factors of traffic flow layer.From the micro point of view,vehicles'lane changing behavior is mostly caused by other vehicles'driving behavior around the current lane.For the certain micro-driving behavior characteristics,the average driving speed will reduce vehicles'lane changing intention,while the vehicles'density will further stimulate vehicles'lane changing behavior.There is a significant cross-layer interaction effect between the micro-vehicle individual and the macro-traffic flow indexes.The multi-layer linear model considering macro and micro influencing factors has higher goodness of fit and model accuracy than the single-level model and the single-layer stochastic utility theory model.The model reflects the interactive behavior of vehicles'lane changing more veritably,and can better explain the nested relationship between vehicle

关 键 词：交通工程 换道意图 多层线性模型 跨层交互 影响因素分析 

分 类 号：U491[交通运输工程—交通运输规划与管理]