交通网络中动态OD矩阵估计模型研究

动态起点-迄点（OD）矩阵为智能交通系统提供所需的基础数据。为了提高交通网络中动态OD矩阵估计的精度,在最小化路段实际交通量和分配交通量之间差值以及最小化历史OD矩阵和待估计OD矩阵之间差值的基础上,将浮动车采集信息融合到OD矩阵估计过程中;在动态网络加载过程中,OD矩阵估计值如果大于加载车辆数会使网络中的定点问题求解发散。为了避免该情况的发生,基于广义最小二乘模型并引入容量约束,提出1种带不等式约束的动态OD矩阵估计方法。仿真实验表明,该方法对于OD矩阵估计的精度较好,优于带滑动窗的广义最小二乘算法。

A Novel Model for Dynamic OD Matrix Estimation in Traffic Network

Dynamic OD matrix provides the basic data for the intelligent transportation system.In order to improve the precision of dynamic OD matrix estimation,the dispersion between the road flows and the assignment flows and the dispersion between the prior OD matrix and the estimated OD matrix are minimized.Simultaneously,the information that is collected by the probe vehicles is fused in the process of the OD matrix estimation.In the dynamic networks loading process,the iterative process of the fixed-point issue for the traffic network will diverge when the estimated value of the OD matrix exceeds the number of the load vehicles.Therefore,the capacity constraint is used in the estimation process.Based on the Generalized Least Square（GLS） model,a dynamic OD matrix estimation algorithm with inequality constraints is proposed.The simulation tests show that the proposed method can get good performance for OD matrix estimation,and the estimation accuracy is much higher than that of the algorithm with the sliding window.