基于NGSIM数据的车辆换道前跟驰模型研究

Improving Car Following Model Before the Vehicle Changes Lanes Based on NGSIM Data

为研究车辆在换道前与车辆在正常行驶时跟驰行为的差异性,从NGSIM数据库中提取了快速路上的543个换道行为和870个非换道行为的跟驰事件并进行特征分析。使用曼-惠特尼U检验(Mann-Whitney U test)验证了换道前车辆的跟驰行为与正常行驶的跟驰行为存在显著差异。选取跟驰车辆的车速和车头间距作为性能指标,其均方根百分比误差之和为目标函数,并将目标车道的前车速度纳入到智能驾驶员模型(IDM)中,构建换道准备智能驾驶员跟驰模型(BLC-IDM),利用遗传算法对BLC-IDM进行参数标定和效果验证。研究结果表明,传统的IDM不适用于换道前车辆的跟驰行为,改进后的BLC-IDM拟合精度提高了20%。BLC-IDM可以更加精准地描述车辆换道前的特殊跟驰行为。

In order to investigate the difference between the following behaviour of vehicles before lane change and vehicles during normal driving,543 following events of lane change and 870 following events of non-lane change behaviour on fast roads were extracted from the NGSIM database and characterised.The Mann-Whitney U test was used to verify that the following behaviour of vehicles before lane change was significantly different from that of normal driving.The speed and headway of the following vehicle were selected as the performance indicators,and the sum of their root mean square percentage errors was used as the objective function.The speed of the vehicle in front of the target lane was incorporated into the Intelligent Driver Model(IDM)to construct the Before Lane-change Intelligent Driver Model(BLC-IDM),and the parameters of the BLC-IDM were calibrated and validated using genetic algorithms.The results show that the conventional IDM is not suitable for the before lane change vehicle following behaviour,and the accuracy of the improved BLC-IDM can be improved by 20%,which can more accurately describe the special following behaviour of the vehicle before lane change.