短期高速公路交通流量预测方法研究
发布时间：2019-02-28 21:30
【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。
[Abstract]:The prediction of short-term traffic flow can predict the development trend of expressway traffic flow in the short time in the future, guide the real-time traffic operation and management of expressway and improve the traffic condition. The traditional short-term forecasting model only reflects some information of traffic flow, which is greatly influenced by the complexity and nonlinearity of traffic components and non-linearity of highway flow data, and the prediction accuracy is low. In order to improve the prediction accuracy of highway short-term traffic flow, an improved traffic forecasting method is proposed based on the periodic characteristics of traffic flow data. First, the periodic component of the flow data is extracted, and then the residual component after removing the periodic component is predicted by using the autoregressive moving average prediction model. Finally, the residual component prediction is accumulated with the periodic component. The final prediction is obtained. Several groups of comparative experiments were carried out to study the effect of different proportion of periodic components on the prediction. When the residual component is negative, the periodic component is modified by the method of increasing and decreasing the offset. The experimental results show that after the periodic component is modified, the accuracy can be improved by extracting the periodic component data, and the higher the energy ratio of the periodic component is, the more obvious the accuracy improvement will be.
【作者单位】： 四川大学电气信息学院;
【基金】：四川省交通科技项目(2013c7-1)
【分类号】：U491.14
[Abstract]:The prediction of short-term traffic flow can predict the development trend of expressway traffic flow in the short time in the future, guide the real-time traffic operation and management of expressway and improve the traffic condition. The traditional short-term forecasting model only reflects some information of traffic flow, which is greatly influenced by the complexity and nonlinearity of traffic components and non-linearity of highway flow data, and the prediction accuracy is low. In order to improve the prediction accuracy of highway short-term traffic flow, an improved traffic forecasting method is proposed based on the periodic characteristics of traffic flow data. First, the periodic component of the flow data is extracted, and then the residual component after removing the periodic component is predicted by using the autoregressive moving average prediction model. Finally, the residual component prediction is accumulated with the periodic component. The final prediction is obtained. Several groups of comparative experiments were carried out to study the effect of different proportion of periodic components on the prediction. When the residual component is negative, the periodic component is modified by the method of increasing and decreasing the offset. The experimental results show that after the periodic component is modified, the accuracy can be improved by extracting the periodic component data, and the higher the energy ratio of the periodic component is, the more obvious the accuracy improvement will be.
【作者单位】： 四川大学电气信息学院;
【基金】：四川省交通科技项目(2013c7-1)
【分类号】：U491.14
【相似文献】
相关期刊论文 前2条
1 杨茂;齐s，
本文编号：2432186
本文链接：https://www.wllwen.com/kejilunwen/daoluqiaoliang/2432186.html
