摘要
文中提出了一种基于北斗高精度定位轨迹数据的车辆换道行为识别模型.该模型提出搜索算法对换道开始时刻点、换道时刻点、换道结束时刻点进行搜索并标定后,增加车辆换道横向偏移距离这一约束条件用以区分车辆蛇形驾驶和车辆换道.利用Spss对模型输出数据进行层次聚类,将驾驶员分为稳重型和活跃型两组类别.结果表明:活跃型驾驶员换道次数较频繁,行驶速度较快,换道时间和横向偏移距离较为随机,多为中间层级驾龄的驾驶员.稳重型驾驶员主动换道行为较少,行驶速度相对不高,换道时间与横向偏移距离表现出明显的驾驶行为习惯特征,多为低驾龄、行车谨慎和高驾龄、操作平稳的人群.
A vehicle lane-changing behavior recognition model based on Beidou high-precision positioning trajectory data was proposed.In this model,a search algorithm was proposed to search and calibrate the lane-changing start time,lane-changing time and lane-changing end time,and then the constraint condition of vehicle lane-changing lateral offset distance was added to distinguish vehicle snake driving from vehicle lane-changing.The model output data was hierarchically clustered by Spss,and the drivers were divided into two groups:stable drivers and active drivers.The results show that active drivers change lanes frequently,travel faster,change lanes at random time and lateral offset distance,and most of them are drivers with intermediate driving experience.Steady drivers have less active lane-changing behavior and relatively low driving speed.The lane-changing time and lateral offset distance show obvious characteristics of driving habits,and most of them are people with low driving experience,cautious driving and high driving experience and stable operation.
作者
徐文洁
赵欣
酆磊
陈曦
XU Wenjie;ZHAO Xin;FENG Lei;CHEN Xi(School of Transportation and Logistics Engineering,Wuhan University of Technology,Wuhan 430063,China)
出处
《武汉理工大学学报（交通科学与工程版）》
2023年第2期239-244,250,共7页
Journal of Wuhan University of Technology（Transportation Science & Engineering）
关键词
高精度定位
换道模型
驾驶行为
轨迹数据
high-precision positioning
lane changing model
driving behavior
trajectory data
