摘要
短时交通流量预测能够推算高速公路未来短时刻内的交通流量的发展动向,指导即时的高速公路交通运营管理并改善交通运行状况。传统短期预测模型只反映交通流量部分信息,受高速公路流量数据的流量成分复杂性和非线性影响较大,预测精确度较低。为了提高高速公路短期交通流量的预测精度,结合交通流量数据中的周期性特征,提出一种改进的流量预测方法。首先提取流量数据的周期分量,然后用自回归滑动平均预测模型对去除周期分量后的残余分量进行预测,最后将得到的残余分量预测值与周期分量进行累加,得到最终的预测值。并进行若干组对比实验研究周期分量比例不同对预测的影响。当残余分量出现负值时,通过增减偏移量的方法对周期分量进行修正。实验表明,修正了周期分量后,提取了周期分量的数据再进行预测,精度能得到提高;周期分量的能量比例越大,精度提升越明显。
In order to enhance prediction precision of short-term highway traffic flow,an improved traffic prediction method is put forward on the basis of the character of the periodicity.First of all,periodic components of flow data are removed from the flow data;secondly,prediction based on residual components by the method of auto regressive moving average are made;finally,the predicted residual components are added to the periodic component to obtain a final predicted value.Removing the periodic component helps highlighting individual characteristics of traffic flow,and reducing interference of noise.Several groups of comparative experiments with different periodic components are done.When residual component turns negative,the offset has to be obtained or abandoned to modify the periodic component and residual component.Experimental results show that the accuracy of traffic prediction can be obviously improved and the greater the proportion of the periodic components,the more precision improved.
出处
《计算机仿真》
北大核心
2017年第3期123-127,共5页
Computer Simulation
基金
四川省交通科技项目(2013c7-1)
关键词
高速公路
短期交通流预测
自回归滑动平均
周期分量
残余分量
Highway
Short-term traffic flow prediction
Auto regressive moving average
Periodic component
Residual component