交通运输系统工程与信息 ›› 2019, Vol. 19 ›› Issue (5): 50-58.

• 智能交通系统与信息技术 • 上一篇    下一篇

基于在线地图交通状态的关键道路动态识别方法

陈华伟a,邵毅明*a, b,敖谷昌a, b ,张惠玲a, b   

  1. 重庆交通大学a. 交通运输学院;b.山地城市交通系统与安全重庆市重点实验室,重庆 400074
  • 收稿日期:2019-03-01 修回日期:2019-05-07 出版日期:2019-10-25 发布日期:2019-10-25
  • 作者简介:陈华伟(1993-),男,海南海口人,博士生.
  • 基金资助:

    国家自然科学基金/ National Natural Science Foundation of China(51508061);山地城市交通系统与安全重点实验室开放基金/Fund of Chongqing Key Lab of Traffic System & Safety in Mountain Cities(2018TSSMC03);重庆市教委科学技术研究项目/Science and Technology Research Program of Chongqing Municipal Education Commission(KJQN201800727).

Dynamic Identification Method of Critical Roads Based on Traffic State of Online Map

CHEN Hua-weia, SHAO Yi-minga, b, AO Gu-changa, b, ZHANG Hui-linga, b   

  1. a. School of Traffic & Transportation; b. Chongqing Key Lab of Traffic System & Safety in Mountain Cities, Chongqing Jiaotong University, Chongqing 400074, China
  • Received:2019-03-01 Revised:2019-05-07 Online:2019-10-25 Published:2019-10-25

摘要:

交通拥堵问题日趋严重,关键道路识别成为了交通领域的研究重点. 以在线地图的交通状态数据为基础,利用时空相关性理论计算道路交通状态的预测值和波动影响值,并通过Moran 散点图划分道路类型,提出了基于在线地图交通状态的关键道路动态识别方法. 首先,调用在线地图开发者平台API 采集路网的交通状态数据,利用集成学习动态预测交通状态;其次,分析交通状态波动的传播结构,以此量化道路对其近邻道路的影响值;然后,结合道路交通状态的预测值和波动影响值划分道路类型,进而识别出关键道路;最后,以实际路网为例,论证方法可行性.

关键词: 交通工程, 关键道路, Moran散点图, 在线地图, 集成学习

Abstract:

Traffic congestion becomes more and more serious, and critical roads identification has become a research focus in the field of transportation. Based on the traffic state data of online map, this paper calculates prediction of traffic state and affection of traffic state wave by using spatial-temporal correlation theory, classifies different road types by using Moran scatterplot, and then proposes the dynamic identification method of critical roads based on traffic state of online map. Firstly, API of developer platform provided by online map is employed to collect the traffic state in road network, to dynamically predict traffic state by using ensemble learning. Then, the propagation structure of traffic state wave is analyzed to quantify the affection of a road on the roads neighbored it. Next, to identify critical roads, different road types are classified according to the prediction of traffic state and the affection of traffic state wave. Finally, taking actual road network as an example, the feasibility of the method is demonstrated.

Key words: traffic engineering, critical roads, Moran scatterplot, online map, ensemble learning

中图分类号: