城市快速路交通流参数估计方法研究

Research on Estimation Method of Traffic Flow Parameters for Urban Expressway

车辆速度、密度、流量关系模型是研究交通流道路通行能力和交通运行状况的重要依据,针对真实道路车辆速度估计,展开经典交通流参数估计解析模型和基于机器学习方法模型的估计效率研究,并提出一种聚类最小二乘方法估计车辆速度。真实交通道路往往存在大量随机因素,而且现场采用的是微波监测器采集数据,导致原始数据存在极大的不确定性,因此,该方法首先对实测数据进行修复、校正和平滑等预处理;为了提高参数估计的效率和准确性,该方法采用k均值聚类对预处理后数据进行聚类;最后采用最小二乘法估计车辆速度。利用实测数据对算法进行测试,结果表明,文章所提出的算法估计效果优于经典解析模型,提高了交通流参数的估计精度,对更加精确的刻画交通流变化趋势有一定的现实意义。

The relationships of the speed,density,and flow are important basis for studying the traffic capacity and traffic conditions of traffic flow.Aiming at the estimation for real vehicle speed of some urban expressway,the estimation efficiency of classical analytical model and a method based on machine learning model were researched.A method named clustering least squares was proposed.There are often a lot of random factors in real traffic roads,and microwave monitor are usually used to collect data on the field,which leads to great uncertainty in raw data.Therefore,firstly,the measured data were preprocessed by using some methods such as repairing,correcting,and smoothing.Secondly,the preprocessed data were used to cluster by k-means.Finally,the vehicle speed was estimated by the least squares.These methods were tested by measured data collected by some urban expressway,and the experimental results show that the proposed algorithm is better than the classical analytical model,it improves the estimation accuracy of traffic flow parameters,and has certain practical significance for more accurately describing the trend of traffic flow.