在对高速公路的起一讫(OD)矩阵进行动态估计时,往往忽略了交通流在运行实际过程中运行时间和车队离散的现象.分析了这一问题对高速公路的OD估计的影响,并通过时空分析引入了交通流“时-空带”的概念,消除了同一股车流在道路下游出现的时间不同步性对于OD估计的不利影响.在上述分析基础上,采用最小二乘理论建立动态OD矩阵估计模型.数值算例表明,该估计模型能够有效地过滤交通流运行时间的影响,提高动态OD矩阵的估计精度,具有较强的实用性.
Most existing dynamic origin-destination (OD) matrix estimation approaches are based on synchronous observed volume. However, when traffic flow runs on a longdistance freeway, the big time lag and platoon dispersion will affect the precision of OD estimation. The paper puts forward the conception of "volume space-time belt" and a new OD estimation methodology is established based on space-time analysis. Traffic volume on space-time diagram is distributed to each "volume belt" and OD matrix is then estimated with the least square method. A practical sample testifies that the new dynamic OD estimation methodology can overcome the defect due to unsynchronized flow and better trace the variation of OD matrix. This character is very important for high simultaneity and accuracy required traffic control system.