摘要
为识别山区公路弯道车辆轮迹集中分布区,实现对轮迹集中分布区内的交通标线和道路结构的建设与养护,基于实地行车状态数据,采用BP神经网络建立了双车道公路弯道左、右转车辆轮迹模型;采用重心聚类法对驶经弯道的车辆进行分类,根据道路参数和占比最大车辆入弯时行驶状态的临界值,建立轮迹模型,计算出弯道左转车辆、右转车辆和左、右转车辆叠加轮迹集中分布区.结果表明:弯道的左、右转车辆叠加轮迹集中分布区内的交通标线范围为-28.54~-24.38 m和-13.20~18.67 m;叠加轮迹集中分布区在道路横断面方向上的分布是-150~50 cm.
Locating the concentrated wheelpath distribution area of turning vehicles on the curve of mountainous highway is the prerequisite for targeted traffic marking and road structure maintenance.Based on the actual driving status data,wheelpath models of left-and right-turning vehicles on the curves of two-lane highway were established based on BP neural network.Vehicles entering the curve were classified by centroid clustering.Then the respective wheelpath distribution areas for left-and right-turning vehicles,and the overlapped wheelpath areas were calculated according to the geometric parameters of the curves and critical running status of the vehicles constituting the largest proportion when making turning.The results show that the overlapped wheelpath areas for left-and right-turning vehicles concentrated on-28.54~-24.38 m and-13.20~18.67 m in traffic markings intervals.The overlapped wheelpath distribution area concentrated in-150~50 cm in the cross section of the roads.
作者
孟欣强
郭建钢
MENG Xinqiang;GUO Jiangang(College of Transportation and Civil Engineering,Fujian Agriculture and Forestry University,Fuzhou,Fujian 350003,China)
出处
《福建农林大学学报(自然科学版)》
CSCD
北大核心
2021年第1期140-144,共5页
Journal of Fujian Agriculture and Forestry University:Natural Science Edition
关键词
公路弯道
轮迹
BP神经网络
聚类分析
highway curve
wheelpath
BP neural network
cluster analysis