摘要
针对高速公路交通拥挤日益严重的现象,通过对收费数据的深层挖掘和高效利用,提出了基于滚动时间序列的行程时间数据合成方法,以此为基础构建了交通拥挤指数,并基于交通拥挤指数的变化特征对拥挤持续时间进行了在线估计;结合收费站布局的时空特征,设计了基本路段和复合路段融合的高速公路交通拥挤自动判别方法.实证分析表明,该方法在判别率提高到96.52%,误判率降低到0.43%的同时,判别时间减少了74%,而且收费数据的获取成本为零.
In view of increasingly serious traffic congestion on freeway, a synthesis method of travel time date was proposed based on the rolling time sequence and charging data, and on which this paper built a traffic congestion index and estimated the duration time according to the changing characteristics of the index. Moreover, taking the spatial and temporal characteristics of toll station layout into account, a method of automatic traffic congestion identification on freeway was designed by merging basic links with composite links. Empirical analysis shows that this method can improve the recognizing rate to 96.52%and reduce the false recognizing rate to 0.43%, at the same time, the recognizing time is declined by 74%, and the cost of charging date is zero.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2014年第12期108-113,共6页
Journal of Harbin Institute of Technology
基金
国家自然科学基金(51278257)
浙江省自然科学基金(LY12F01013)
高等学校博士学科点专项科研基金(20110061110034)
关键词
交通工程
数据合成
交通拥挤
自动判别
收费数据
traffic engineering
data synthesis
traffic congestion
automatic identification
charging date
