摘要
为排除驾驶员在换道过程中存在的交通安全隐患,基于驾驶员操纵特性和交通环境状态分析,提出了一种能有效预测驾驶员换道行为的方法。依托自然道路条件下的实车实验,将驾驶员视觉特性与数据分析相结合,确定了意图时窗宽度为5 s。依据在车道保持与意图换道两阶段的操纵特性和交通环境等参数差异化分析,构建了换道行为预测的表征参数体系。引入二元Logistic模型,确定各参数的回归系数,并进行换道行为预测。研究结果表明:该模型可至少提前2. 5 s预测出换道行为;且换道行为起始时刻的预测精度为96. 34%。与基于视觉特性和转向灯状态预测换道行为相比,具有更高的准确率和更优的时序性。
In order to eliminate potential traffic safety during the lane-changing process,based on the analysis of handling characteristics and traffic environment states,a new method to identify drivers’ lane-changing behavior was proposed. Relying on the real-world experiments,the drivers’ visual characteristics and data analysis were combined to determine the intent window width of 5 s. Depending on the difference analysis of the handling characteristics and traffic environment between the lane-keeping and lane-changing intent stage,the characterization index system for predict lane-changing behavior was constructed. The two element Logistic model was introduced to determined the regression coefficient of each parameter and predict the lane-changing behavior. The research result show that the model may predict drivers’ lane-changing behavior for at least 2. 5 s in advance,and the prediction accuracy of the start time of the lane-changing behavior is 96. 34%. The accuracy and time series characteristics of the model are superior to the use of turn signals and visual characteristics in predicting lane-changing behaviour.
作者
李青林
彭金栓
付锐
徐磊
方媛
LI Qing-lin;PENG Jin-shuan;FU Rui;XU Lei;FANG Yuan(School of Traffic & Transportation, Chongqing Jiaotong University, Chongqing 400074, China;School of Automobile, Chang an University, Xi an 710064, China)
出处
《科学技术与工程》
北大核心
2019年第12期368-377,共10页
Science Technology and Engineering
基金
国家自然科学基金(61503049)
重庆市自然科学基金(cstc2018jcyjAX0288)
重庆市博士后研究人员科研项目特别资助项目(Xm2016056)
重庆市教委科学技术研究项目(KJ1600540)
山地城市交通系统与安全重庆市重点实验室开放基金(KTSS201602)资助
关键词
意图时窗
换道行为
预测指标
LOGISTIC模型
时序性
intent time window
lane-changing behavior
predictive index
Logistic model
time series characteristics
