基于CHMM的信号交叉口车流转向比例估计研究  

Coupled Hidden Markov Model Based Turning Movement Proportions Estimation at Signalized Intersections

作　　者：盛东雪 曹鹏 SHENG Dongxue;CAO Peng(School of Transportation and Logistics,Southwest Jiaotong University,Chengdu 611756,China)

机构地区：[1]西南交通大学交通运输与物流学院,四川成都611756

出　　处：《综合运输》2021年第11期81-87,共7页China Transportation Review

摘　　要：为估计城市干道上信号交叉口处交通流的转向比例,提出利用浮动车轨迹数据,基于耦合隐马尔可夫模型(CHMM)的交叉口转向比例估计方法。基于交通流理论与CHMM建模分析城市干道的交通状态时空演变过程,使用期望最大化(EM)算法迭代求解交叉口各进口道交通流的转向比例。在VISSIM微观仿真软件中搭建含有两个信号交叉口的城市干道网络,运行仿真以获取模拟不同渗透率下的浮动车轨迹数据,并将估计的转向比例结果与仿真设置的真实值对比。研究表明,随着浮动车渗透率增加估计结果的MAPE持续下降,且当渗透率超过14%时,MAPE降至0.30以下,验证了所提出方法的准确性。In order to estimate the turning movement proportions at signalized intersections on urban arterial roads,a method which is based on coupled hidden markov model(CHMM)and using the trajectory data of probe vehicles was proposed.Based on traffic flow theory and CHMM,the spatio-temporal evolution process of traffic state of urban arterial roads was analyzed,and the expectation maximization(EM)algorithm was used to solve the turning movement proportions of each entrance road at the intersection iteratively.An urban arterial road network with two signalized intersections was built in VISSIM microsimulation software.The simulation was run to obtain the trajectory data of the simulated probe vehicles with different penetration rates,and the estimated turning movement proportions were compared with the real value.The results show that the MAPE of the estimated results continues to decrease with the increase of the penetration rate,and when the penetration rate exceeds 14%,the MAPE drops to less than 0.30,which verifies the accuracy of the proposed method.

关 键 词：信号交叉口 转向比例 CHMM 粒子滤波算法 EM算法 

分 类 号：U491[交通运输工程—交通运输规划与管理]