电子学报 ›› 2015, Vol. 43 ›› Issue (12): 2491-2496.DOI: 10.3969/j.issn.0372-2112.2015.12.022

• 学术论文 • 上一篇    下一篇

基于N阶近邻路网的车辆行程时间估计模型

章登义, 欧阳黜霏, 吴文李   

  1. 武汉大学计算机学院, 湖北武汉 430072
  • 收稿日期:2015-01-28 修回日期:2015-07-30 出版日期:2015-12-25
    • 通讯作者:
    • 欧阳黜霏
    • 作者简介:
    • 章登义 男,1965年5月出生于湖北省荆州市.现为武汉大学计算机学院教授、博士生导师
    • 基金资助:
    • 国家自然科学基金 (No.60903035,No.41001296); 国家863高技术研究发展计划 ( (No.2013AA12A301)

N-order Neighbor Road Network Based Vehicle Travel Time Estimation Model

ZHANG Deng-yi, OUYANG Chu-fei, WU Wen-li   

  1. School of Computer, Wuhan University, Wuhan, Hubei 430072, China
  • Received:2015-01-28 Revised:2015-07-30 Online:2015-12-25 Published:2015-12-25
    • Supported by:
    • National Natural Science Foundation of China (No.60903035, No.41001296); National High Technology Research and Development Program of China  (863 Program) (No.2013AA12A301)

摘要:

车联网的提出为智能交通的研究提供了新的交通信息收集技术.针对短时交通中车辆的路网行程时间估计问题,提出了基于N阶近邻的隐马尔科夫模型,利用马尔科夫性质来解决道路行程时间的前后关联性问题,同时考虑不同道路的异构性构建了N阶近邻路网模型来模拟路网间的交互影响.针对短时交通中实时数据更新的问题,提出基于道路关联性算法,并结合车联网的采集技术给出了迭代更新模型的方法.实验表明,本文提出的方法在短时交通车辆行程时间预测中精度较高,能够在车辆行进中做出实时预测.

关键词: 行程时间预测, 隐马尔科夫模型, 聚类

Abstract:

The development of Internet of vehicles provides a new traffic information collection technique for the study of intelligent transportation.In this article,we propose an N-order hidden Markov model to approach the vehicle travel time prediction problem,utilizing the Markov nature to model the internship of road network.We also promote an N-order neighbor road network to address the heterogeneity of road.A non-trivia update algorithm is applied to handle the real time data approaching issue.We also prove the temporality of the N-order hidden Markov model in travel time prediction.Experimental results on authentic data indicate the effectiveness and accuracy of this approach.

Key words: travel time prediction, hidden Markov model, cluster

中图分类号: