摘要
为提高公交行程时间预测与信息发布的准确性,借助显著性分析确定影响可靠性预测的主要因素.首先,基于公交GPS数据,采用地图匹配算法建立站点区间行程时间计算方法;其次,针对11组不同路段站点之间的区间行程时间数据,通过拟合优度检验筛选最佳分布模型,并利用最大似然估计获取最优分布模型参数;最后,建立公交行程时间可靠性评价指标体系,分析交通条件、道路条件、采样间隔与行程时间波动指数、延误指数的相关关系.结果表明:三元高斯混合分布模型能以100%的接受率最优地拟合公交行程时间数据,站点区间长度、公交小时流量、采样间隔与行程时间可靠性存在相关关系,而交叉口相对位置则为非关键影响因素.
In order to improve the accuracy of public travel time forecasting and information dissemination,the factors that affect the reliability prediction are determined by means of significance analysis. First,based on GPS（ global positioning system） data of public transport,the calculation method of travel time betw een stations is built up by the map-matching algorithm. Then,the optimum distribution model is built up by using the test of fit goodness with 11 groups of travel time data of different road sections,and the parameters of the optimal model are obtained by the maximum likelihood estimation. Finally,the evaluation index system of the bus travel time reliability is established to analyze the correlation betw een the traffic condition,road condition,sampling interval and the fluctuation index as well as the delay index. The results show that the Gaussian mixture distribution model can fit the public transport travel time best with the acceptance rate of 100%. The length betw een tw o stations,public traffic per hour and sampling interval are correlative to the travel time reliability,while the location of junctions is the non-critical factor.
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2015年第2期404-412,共9页
Journal of Southeast University：Natural Science Edition
基金
国家重点基础研究发展计划(973计划)资助项目(2012CB725402)
