基于数据相似性的路段短期交通流预测方法研究  

Short term traffic flow prediction method based on data similarity

作　　者：于德新 常丽君 郭海波 田广东 YU De-xin;CHANG Li-jun;GUO Hai-bo;TIAN Guang-dong(School of traffic engineering,Jilin Institute of architecture and technology,Changchun 130114,China;School of transportation,Jilin university,Changchun 130022,China;School of mechanical engineering,Shandong university,Jinan 250061,China)

机构地区：[1]吉林建筑科技学院交通工程学院,长春130114 [2]吉林大学交通学院,长春130022 [3]山东大学机械工程学院,济南250061

出　　处：《吉林建筑大学学报》2022年第5期28-32,44,共6页Journal of Jilin Jianzhu University

基　　金：吉林省科技发展计划项目(20190101023JH);吉林省教育厅科学技术研究项目(JJKH20211368KJ)。

摘　　要：机动车保有量增加引发的环境污染与交通拥堵日趋严重,人们对于缓解交通拥堵的交通流预测方法研究逐渐重视.本文对厦门市城市道路的RFID(Radio frequency identification)数据预处理后进行相似性分析,选取相似性系数和不稳定系数为指标来表征路段交通流的相似度,并判断阈值以及对应交通状态下的路段交通转移量,最终以数据相似性为基础,建立交通流预测模型.选取10条城市道路RFID数据验证,其平均绝对误差为0.061,证明本文提出模型的可行性.With the problem of traffic congestion and environmental pollution caused by the increase of motor vehicle ownership becoming more serious,people pay more attention to the research on traffic flow prediction methods to alleviate traffic congestion.Based on the RFID(Radio Frequency Identification)data of urban roads in Xiamen,analyze the similarity after preprocessing,use the similarity coefficient and instability coefficient index to characterize the similarity degree of road traffic flow,judge the threshold and the road traffic transfer volume under the corresponding traffic state,and finally establish the traffic flow prediction model based on data similarity.The RFID data of 10 urban roads are selected to verify that the prediction results are more accurate,which proves the effectiveness of the traffic flow prediction model proposed in this paper.

关 键 词：交通流预测 RFID 数据相似性 流量转移率 

分 类 号：U491.14[交通运输工程—交通运输规划与管理]