摘要
依托卡口检测数据,提出路径拼接模型以利用大量短路径行程时间对城市道路中的长路径行程时间进行估计。该模型提出子路径进行拼接时要满足的两个条件:存在公共结点及交通状态相近;在此基础上将目标路径按照整体方差最小的原则拆分为由多个子路径组成的拼接方案。基于提出的路径交通流基本图划分子路径的交通状态,使用Burr分布拟合不同交通状态下的子路径行程时间得到行程时间分布函数;将给定的目标行程时间分配至拼接方案中的子路径使得子路径行程时间之和等于目标行程时间,累加所有时间分配方案对应的概率得到目标路径的行程时间概率,计算多个目标行程时间对应的概率得到目标路径的行程时间概率分布。实验结果显示:本文提出的路径行程时间估计方法得到的路径行程时间均值与真实轨迹均值的误差为3.04%;得到的行程时间分布与真实行程时间分布的JS散度为0.05。路径行程时间估计值真实可靠,可为后续研究提供数据基础。
In this paper,we propose a path splicing model to estimate the long path travel time on urban roads using the automatic vehicle identification(AVI)data.The model requires two conditions to be met when splicing the sub-paths:public nodes exist and the traffic status is similar.A target path is split into a splicing scheme composed of several subpaths whose variance is the least.Based on the proposed traffic state,the travel time distributions are obtained by fitting the travel time under different states using the Burr distribution.Next,the method assigns target travel time to the sub-paths,thus,the travel time probability of the target path is the accumulation of the probability of all time assignments.The study shows that the error between the average path travel time of the proposed method and the real path is 3.04%,and the JS divergence between the travel time distribution of the method and the real path is 0.05.The estimated path travel time is reliable,which can provide a data basis for subsequent research.
作者
李烨焘
黄敏
黄春婷
张小兰
LI Yetao;HUANG Min;HUANG Chunting;ZHANG Xiaolan(School of Intelligent Systems Engineering/Guangdong Provincial Key Laboratory of Intelligent Transportation System,Sun Yat-sen University,Shenzhen 518107,China;Guangdong Polytechnic of Industry and Commerce,Guangzhou 510510,China)
出处
《中山大学学报（自然科学版）（中英文）》
CAS
CSCD
北大核心
2023年第6期80-88,共9页
Acta Scientiarum Naturalium Universitatis Sunyatseni
基金
国家重点研发计划(2020YFB1600400)。
关键词
路径行程时间
路径拼接
交通状态
行程时间分布
path travel time
path splicing
traffic status
distribution of travel time
作者简介
李烨焘(1997年生),男,研究方向:智能交通系统,E-mail:liyd33@mail2.sysu.edu.cn;通信作者:黄敏(1975年生),女,研究方向:智能交通系统,E-mail:huangm7@mail.sysu.edu.cn。
