城市快速路短时交通量预测方法研究

Research on Short-term Traffic Volume Prediction Methods for Urban Expressways

摘要:

随着我国城市规模的快速发展，大型城市中的快速路已成为人们主要的出行选择，而早晚高峰的短时交通量暴增可能引发交通拥堵等问题，采用合理的方法对其进行预测，可实现有效的交通诱导、降低交通事故发生率和缩短延误时间等。根据交通流的离散性，将单日时间分成3个时间段，采用灰色系统理论对每个时间段分别建立5、10、15 min等步长的交通量预测GM（1，1）及其残差模型。将该方法应用于广州市广园快速路的短时交通量预测，分析不同时间段内多种步长预测方式的误差结果，通过残差率对比分别确定模型的适应时间段。将预测结果与传统线性回归的预测方法进行对比，验证模型的准确性。

Abstract:

With the rapid development of urban scale in China, expressway has become the main way of travel in large cities, and the short-term surge of traffic volume in the morning and evening peak may lead to traffic congestion and other problems. The effective traffic guidance can be realized, the traffic accidents can be reduced and the delay time can be shortened by using reasonable method to predict it. According to the dispersion of traffic flow, the daily time is divided into three periods. The traffic prediction GM（1，1）and its residual models with steps of 5 min, 10 min and 15 min for each period are established by using the grey system theory. This method is applied to the short-term traffic prediction of Guangyuan Expressway in Guangzhou. The error results of various step size prediction methods in different time periods are analyzed. The adaptation time periods of model are determined by comparing the residual rate. The prediction results are compared with the traditional linear regression prediction method to verify the accuracy of the model.