基于收费数据的高速公路行程时间可靠性估计模型

Reliability Estimation Model of Freeway Travel Time Based on Toll Data

高速公路收费数据易于获取,且蕴含多种交通信息。本文利用收费数据分析高速公路客货车行程时间分布,提出客货车行程时间分布的差异,分别构建客货车行程时间可靠性估计模型,并对模型进行试验分析。利用MATLAB对直接求取的客货车行程时间和模型求取的客货车行程时间进行正态分布和对数正态分布曲线拟合,结果表明:客货车的行程时间分布均符合正态分布;利用本文模型求取的行程时间分布比直接使用出入口收费站时间差求取的行程时间分布具有更小的SSE和RMSE以及更大的R-square。最后,对客货车行程时间的可靠性区间,即置信区间进行比较,结果表明:在同一置信水平下,利用本文模型求解的行程时间具有更小的行程时间区间分布,即更好的稳定性。

Mass freeway toll data contain various traffic information and are readily available.This paper analyzes the difference of travel time distributions of freeway passenger car and freight car and constructs different reliability estimation models for them.Then the paper fit above data with normal and lognormal distributions using MATLAB.The result shows that it＇s acceptable to analyze the distributional characteristics of travel time for passenger car and freight car with normal distribution.The SSE,RMSE,and R-square of travel time distribution are better by model method than those by direct method.Finally,the reliability interval of both types of cars are compared and the result shows that it is more stable with model method under the same probability.