Abstract:

A model is presented for the urban real-time vehicle velocity estimation problem. It uses traffic flow attribute matching to estimate vehicle velocity and uses index on traffic flow attribute to guarantee real-time estimation. The model is validated by the traffic data collected on main streets in Hangzhou, comparing with data regression method and principle curve model. Result shows the practicality and superiority upon other models.

Key words: transportation engineering, vehicle velocity estimation, traffic flow attribute matching, k-nearest neighbor method, real-time

摘要：

针对城市道路的实时行车速度估计问题，提出一种行车速度实时估计模型，采用基于城市交通流特征匹配方法，通过对交通流特征建立有效索引保证行车速度估计的实时性。利用在杭州市区部分主干道采集的数据作为测试集，与数据回归法以及主曲线建模法进行对比，结果表明，本模型在行车速度估计方面具有实用性和优越性。

关键词: 交通工程, 行车速度估计, 交通流特征匹配, k近邻方法, 实时性

CLC Number: 

• U491