Since it is difficult to observe the passenger flow in different directions among the mixed cross-flow within the urban rail transit station in real time, this paper proposed an analytical method of passenger flow in the station based on multi-source monitoring data. Firstly, by using the passenger flow monitoring data from the automatic fare collection system, intelligent video, and train weighing systems and considering the characteristics and spatial-temporal relationship of passenger flows, a model for estimating passenger flow in different directions within the station was constructed. Then, based on the computational graph structure, the model was abstracted into a multi-layer passenger flow network with spatial-temporal dimensions, and the forward propagation and backward propagation techniques were introduced to solve the established model through iterations. Finally, a case study of a transfer station in G City was conducted. The results indicate that the proposed method can be used to resolve the passenger flow in different directions from multi-source data by time periods. The mean absolute percentage error between the estimated passenger flow and the actual monitoring for 72 time periods in a day does not exceed 10.24%, and the average is around 0.41%.
随着我国大城市轨道交通路网规模扩大、客流量增长,智能化成为城市轨道交通运输组织的发展趋势之一,及时准确地感知车站不同流向(由产生点与消失点定义)的分方向客流(以下简称“分向客流”)的实时分布状况是重要支撑条件。但由于网络拓扑、站内设施布局与客流流线日渐复杂,不同来源与去向的若干分向客流在站内多个地点进行混合、交叉与分流。利用自动售检票(Automatic Fare Collection,AFC)系统、列车称重系统、智能视频等设备仅能够对进出站及交叉分合后形成的客流进行感知,而其中的各个分向客流却难以被直接观测,成为当前客流状态感知的难题。
利用Python3.8.2进行软件开发实现3.2所述算法。选取该站某工作日,以各个时段进站/出站闸机组客流量、来向及去向区间断面客流量、站台与站厅间楼/扶梯通行客流量、车站整体聚集客流量等客流监测信息为输入,运用算法软件迭代求解各个时段的各决策变量。在每个时段,首先对各层节点之间权重,,进行初始化,即先对权重进行随机赋值,再基于约束条件公式⑺与公式⑻对权重进行归一化处理。之后算法共有3个关键步骤:①基于正向传播方式,将到站客流按照各比例和权重依次分布至客流分向、路径、监测点位以及车站整体节点上,得到客流估计值,并借助客流监测信息计算目标函数,;②基于反向传播方式,计算目标函数关于不同层节点间权重的一阶偏导数;③运用共轭梯度法对决策变量进行更新计算,使得目标函数值进一步降低。在判断算法是否停止迭代时,充分小正数取为0.000 1、容忍临界值取为36。在经过339 875次迭代、计算时长3 h 36 min 18 s后,各时段目标值都趋于稳定。前5个时段模型目标函数迭代收敛曲线图如图4所示,各时段目标值分别在25,50,800,1 000,3 000次迭代后,稳定在0.010,0.260,13.155,31.011,12.677左右。
CIPRIANIE, GEMMAA, MANNINIL,et al. Traffic Demand Estimation Using Path Information from Bluetooth Data[J]. Transportation Research Part C:Emerging Technologies,2021,133:103443.
[2]
GUJ J, JIANGZ B, SUNY S,et al. Spatio-Temporal Trajectory Estimation Based on Incomplete Wi-Fi Probe Data in Urban Rail Transit Network[J]. Knowledge-Based Systems,2021,211:106528.
[3]
DANALETA, FAROOQB, BIERLAIREM. A Bayesian Approach to Detect Pedestrian Destination-Sequences from WiFi Signatures[J]. Transportation Research Part C:Emerging Technologies,2014,44:146-170.
LAIJianhui, CHENYanyan, ZHONGYuan,et al. Travel Route Identification Method of Subway Passengers Based on Mobile Phone Location Data[J]. Journal of Computer Applications,2013,33(2):583-586.
LIANGPeng, LIUXiaoyong, HAOGang,et al. Pedestrian Movement Monitoring System and Implementation Based on Face Recognition and RFID Recognition[J]. Journal of Guangdong Polytechnic Normal University,2015,36(11):21-25.
[8]
闫秋芳. 基于行人再识别的轨迹重现[J]. 现代计算机,2020(4):51-54.
[9]
YANQiufang. Track re-Creation Based on re-Identification[J]. Modern Computer,2020(4):51-54.
[10]
吴正阳. 城市轨道交通网络客流分配理论与控制技术研究[D]. 成都:西南交通大学,2018.
[11]
刘卫松. 基于时空网络的城市轨道交通客流分配研究[D]. 成都:西南交通大学,2018.
[12]
HAMDOUCHY, SZETOW Y, JIANGY. A New Schedule-Based Transit Assignment Model with Travel Strategies and Supply Uncertainties[J]. Transportation Research Part B:Methodological,2014,67:35-67.
HASHEMIH, ABDELGHANYK F. Real-Time Traffic Network State Estimation and Prediction with Decision Support Capabilities:Application to Integrated Corridor Management[J]. Transportation Research Part C:Emerging Technologies,2016,73:128-146.
LEEJ Y S, LAMW H K, WONGS C. Pedestrian Simulation Model for Hong Kong Underground Stations[C]//ITSC 2001.2001 IEEE Intelligent Transportation Systems. Proceedings. Oakland,CA,USA. IEEE,2001:554-558.
ZUOJing, YUZhao. Real-Time Train Passenger Flow Detection Algorithm Based on Convolutional Neural Network[J]. Journal of Railway Science and Engineering,2023,20(3):836-845.
PENGYufei, JIANGXi. Estimation Method of Urban Rail Transit Passenger Route Based on Multi-Source Detection Data[J]. Journal of Beijing Jiaotong University,2023,47(3):96-102.
[24]
SHANGP, YANGL Y, YAOY,et al. Integrated Optimization Model for Hierarchical Service Network Design and Passenger Assignment in an Urban Rail Transit Network:A Lagrangian Duality Reformulation and an Iterative Layered Optimization Framework Based on Forward-Passing and Backpropagation[J]. Transportation Research Part C:Emerging Technologies,2022,144:103877.
[25]
WUX, GUOJ F, XIANK,et al. Hierarchical Travel Demand Estimation Using Multiple Data Sources:A Forward and Backward Propagation Algorithmic Framework on a Layered Computational Graph[J]. Transportation Research Part C:Emerging Technologies,2018,96:321-346.