摘要
平面交叉口多相位自适应信号控制需要依据其进口道实时OD流量设置配时参数，以适应变化的交通需求．为此，笔者建立了描述平面交叉口动态交通需求的多变量动态线性模型及其递推算法．该模型基于交叉口进出口断面的流量观测数据，以进口道的分流概率矩阵为状态参数，利用贝叶斯预测方法，对时变OD矩阵进行联线估计．对一特定的T型交叉口交通状态的计算机仿真表明该模型和算法具有较好的精度，能用于交叉口多相位自适应信号控制．
In order to satisfy the time-changing traffic demands at intersections, multiphase adaptive controllingneeds to be dependent on the real-time OD flows of various entries. Therefore, a multivariate dynamic linear modeland its recursive algorithm which is suitable for describing the drnamic traffic patterns at intersections are developed in this paper. Using the Bayesian forecasting method, this model, whose state vectors are OD proportion matrices,canprovide the on-line estimations of time-dependent OD matrices, based on the traffic flow observations of all entriesand exits,A simulation experiment for the traffic patterns at a specified T-intersection is also performed in the lastsection.
出处
《同济大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
1998年第2期172-175,共4页
Journal of Tongji University:Natural Science
关键词
动态线性模型
交通流
时变OD流量
交叉口
Multivariate dynamic linear model
The-dependent OD flow
Recursive estimation
