高速公路交通流量预测模型仿真

摘要

Traffic flow forecasting is essential for highway managers to make decisions,but because of the characteristics of suddenness,time-varying and nonlinearity of short-term traffic flow,the traditional prediction method is of low prediction precision and poor adaptability.In order to improve the accuracy of short-term traffic flow prediction,a time series forecasting method based on Empirical Mode Decomposition (EMD) and Wiener filtering is proposed.First of all,the traffic flow data was decomposed into several intrinsic mode components by EMD.Signal dominant modes were predicted using autoregressive moving average (ARMA) model.The noise dominant modes were denoised with Wiener filter,and predicted using EMD and ARMA model.Finally,the prediction results of each part were combined to obtain the final results.According to the prediction of the traffic flow data in Chengdu,Sichuan,using this method,the proposed method is more accurate than the traditional single model and is an effective method for short-term traffic flow forecasting.%交通流量预测对于高速公路管理进行决策至关重要,但是由于短时交通流量有较强的突发性、时变性和非线性,传统的预测方法预测精度低,适应能力差.为了提高短时交通流量预测精度,提出一种基于经验模态分解(EMD)和维纳滤波预处理的时间序列预测方法.采用EMD将交通流量数据分解成多个本征模态分量,信号主导模态利用自回归滑动平均(ARMA)模型进行预测,而噪声主导模态采用维纳滤波进行去噪处理后再建立模型预测,最后将各部分预测结果线性组合得到最终预测结果.利用上述方法对四川成都的交通流量数据进行预测,结果表明选用的方法比传统单一预测模型有更高的预测精度.