基于微观仿真的快速路短时交通流预测研究

Research on Short-term Traffic Flow Prediction Based on Microcosmic Simulation

随着预测时间跨度的缩短,交通流量的变化显示出越来越强的不确定性,使得一般方法的预测精度大大降低。为了能够及时准确地进行交通流短时预测,提出了一种以城市快速路为研究对象的以微观仿真技术为基础的短时交通流预测的方法。该方法以实时的快速路交通流数据为基础,应用M3分布模型对车辆进行初始分布,基于约束卡尔曼滤波进行OD矩阵估计,利用微观交通仿真的方法对快速路的未来交通流进行预测。经过实测数据验证,算法能对交通参数给出较好的预测结果。

The accuracy of common forecasting method is very low due to the in-veracity of traffic flow changing as the predicting period gets shorten. In order to forecast the traffic flow in short term, taken city expressways as objects, one kind of short-term traffic flow prediction measure based on microcosmic-simulation technology was put forward This method is based on real time expressway traffic data. All the cars were distributed with M3 Model at the beginning. The future expressway traffic flow was predicted in use of Microcosmic Traffic Simulation method and OD matrix estimation based on constrained Kalman filter. The better predict outcome has been approved by practical test.