面向车联网的高速路OD矩阵估计模型  

Origin-Destination Matrix Estimation Model for Freeway Oriented Internet of Vehicles

作　　者：汪成亮[1] 张晨[1] 黄文龙[2] 

机构地区：[1]重庆大学计算机学院,重庆400044 [2]华为技术有限公司香农实验室,广东深圳518129

出　　处：《西南交通大学学报》2013年第6期1078-1083,共6页Journal of Southwest Jiaotong University

基　　金：国家自然科学基金资助项目(61004112);中央高校基本科研业务费专项资金资助项目(CDJZR12180006)

摘　　要：针对高速路OD矩阵的采集与估计,提出采用RFID技术的车联网进行部署,建立有缺失数据的OD矩阵采集方法.用拉格朗日乘数法求解OD矩阵元素概率的极大似然估计,给出基于EM算法的OD矩阵估计算法.通过9×12节点规模的仿真实验结果表明,提出的算法相比模拟采集准确性更高,真实数据的均方根误差降低了77.7%,验证了算法流程和模型的有效性.The internet of vehicles with radio frequency identification technology（ RFID） was applied to capturing an OD matrix with missing data. In the OD matrix estimation model,Lagrange multiplier method was used to solve the maximum likelihood estimation for the probability of OD matrix elements. The algorithm of estimating OD matrix based on the expectation maximization algorithm was developed. Experimental results with 9 × 12 nodes shows that the proposed algorithm is more accurate than the simulation acquisition,and the root mean square error of real data is reduced by 77. 7%,which validates the flow of the algorithm and the effectiveness of the models.

关 键 词：无线射频识别 车联网 EM算法 OD矩阵估计 

分 类 号：U491.1[交通运输工程—交通运输规划与管理]