城市轨道交通网络短时客流OD估计模型

基于状态空间方法构建适用于城市轨道交通网络的短时客流OD(origindestination)估计模型.利用自动售检票数据分析得到OD间乘客行程时间分布特征,构建基于行程时间分布的客流到达系数,以此建立OD流与车站进出站客流间相互关系,并以车站客流分离率为状态变量构建结构化OD矩阵估计状态空间模型.以北京市轨道交通为对象进行案例分析,结果表明,当估计时段长度为15 min时,估计平均相对误差为35.5%;为30 min时,估计平均相对误差为20.4%;为60 min时,估计平均相对误差为16.3%.所构建模型能能有效解决城市轨道交通短时客流估计问题,具有一定的实用性.

Short-time Passenger Flow Origin-destination Estimation Model for Urban Rail Transit Network

A short- time passenger flow origin- destination matrix estimation model based on state- space approach is proposed for urban rail transit network. An arrive coefficient is specially defined for establishing the relationship between OD flows and in-and out-flows at stations, using the passengers'' travel time distribution through statistical analysis of historical automatic fare collection records, then a structured statespace model using station passenger flow split rate as the state variable is proposed. Finally, a numerical example of Beijing subway network is made. The results show that, when the estimation time interval is 15 minutes, the average relative deviations is 35.5%; if the time interval is 30 minutes, the relative deviations is 20.4%; if the time interval is 60 minutes, the relative deviations is 16.3%. Case study validates that the model meets the request of short-time passenger flow estimation for large-scale urban rail transit network and has strong practicability.