Path Travel Time Estimating Method by Incomplete Traffic Data
Path Travel Time Estimating Method by Incomplete Traffic Data
复制标题
DOI：
10.1007/s13177-018-0168-4
复制
发表时间：
2020-01-01
影响因子：
2.1
通讯作者：
Uchida, Kenetsu
中科院分区：
文献类型：
--
作者：
Tani, Ryuichi;Owada, Takashi;Uchida, Kenetsu
本研究提出一种互补使用交通计数器与探测车资料来估计随机路径行程时间的方法。通过最大似然估计法对交通计数器数据进行估计，得到了路网中各O-D对的随机交通需求。随机路径旅行时间被视为一个先验的多变量路径旅行时间分布。估计的随机路径行程时间更新应用贝叶斯推理使用观察到的探测车数据。更新后的路径旅行时间可以看作是一个后验多元路径旅行时间分布。数值实验演示了路径行程时间的推理过程，验证了我们提出的模型。
This study proposes a method estimating stochastic path travel times by using both traffic counter data and probe car data complementarily. A stochastic traffic demand for each O-D pair in a road network is estimated by maximum likelihood estimation with respect to traffic counter data. Stochastic path travel times are addressed as a prior multivariate path travel time distribution. The estimated stochastic path travel time is updated by applying Bayesian inference using observed probe car data. The updated path travel time can be regarded as a posterior multivariate path travel time distribution. Numerical experiments demonstrate inference processes of path travel time and verify our proposed model.