交通运输系统工程与信息

交通运输系统工程与信息 ›› 2006, Vol. 6 ›› Issue (6): 33-40 .

• 智能交通系统与信息技术 • 上一篇    下一篇

“潮汐式”交通中可变通道的通行方向切换算法的研究与应用

宫晓燕1，康胜2   

  1. 1．中国科学院自动化研究所，复杂系统与智能科学重点实验室，北京100080； 2．上海市电气自动化设计研究所有限公司，上海200023
  • 收稿日期:2006-06-16 修回日期:1900-01-01 出版日期:2006-12-20 发布日期:2006-12-20

Study and Application of Traffic Direction Changing Algorithm for Urban Tide Traffic Situation

GONG Xiao-yan1, KANG Sheng2   

  1. 1．Key Lab of Intelligence Sciences and Complexity Systems，Institute Of Automation，Chinese Academy of Sciences，Beijing 100080, China； 2．Shanghai Electrical Automation D&R Institute Ltd.，Shanghai 200023，China
  • Received:2006-06-16 Revised:1900-01-01 Online:2006-12-20 Published:2006-12-20

PDF

摘要： 目前在很多大城市，由于居住区和工作区分布得过于集中，城市的多条主干道出现“潮汐式”交通的现象。本文以上海市外环隧道潮汐交通控制项目为背景，在分析外环隧道两端入口交通数据的基础上，提出基于有序样本聚类和非参数回归交通流预测的隧道可变通道的通行方向切换算法，并最终形成一套潮汐交通控制软件并在外环隧道实际应用。它可根据隧道两端入口处实时变化的交通流，通过短时预测未来的交通流，实时调整可变通道的通行方向，达到充分利用各条车道的目的。现场运行的用户报告表明：本算法有效地缓解了由潮汐交通所带来的隧道单向拥挤，使得隧道的日通行能力提高了3O％，高峰期的通行能力提高了lO％，隧道两端入口的排队长度减少了40％．

关键词: “潮汐式”交通, 短时流量预测, 非参数回归, 有序样本聚类

Abstract: Recently in some large cities，because of inappropriate layout of living area and working area，tide traffic phenomenon occurs in some main roads and thus has become an important reason for traffic congestion．especially during rush hours．So in this paper, based on Tide Traffic Control for in Shanghai Outlines Cross-River Tunnel project，a new algorithm based on ordinal samples cluster method and non parametric regression method is proposed．This algorithm can adjust the traffic direction of direction-changeable lanes automatically and timely by analysis of current traffic flow and short-term forecasted traffic flow of two tunnel entrances．so that all lanes in the turnnel could be used to the fullest．Field tests and user reports show efficiency of this algorithm by 30％ increase of average traffic capability, 10％ increase of rush hour traffic capability and 40％ decrease of average queue length．

Key words: tide traffic, short term traffic forecasting, non parametric regression, ordinal samples cluster