车头时距分布函数的验证、分析与选择  

Verification,Analysis and Selection of Distribution Function of Headways

作　　者：段力[1,2] 过秀成[1] 姚崇富 

机构地区：[1]东南大学交通学院,江苏南京210096 [2]华中科技大学土木工程与力学学院,湖北武汉430074

出　　处：《公路交通科技》2014年第5期147-152,共6页Journal of Highway and Transportation Research and Development

基　　金：国家自然科学基金项目(51208222);江苏省交通科学研究计划项目(2011-Y-30)

摘　　要：采用30组高速公路实测数据，对常见的车头时距分布函数进行验证。为明确其适用性，分析了分布函数的特征，并提出选择策略。研究结果表明负指数分布能较好地反映车辆到达的随机性，且较为简单，当车道交通量小于250veh／h时优先选用。移位负指数分布通过最小车头时距的合理取值减小了拟合误差，而爱尔朗分布的阶数必须取整数，扩大了误差，当交通量位于250～750veh／h时，应优先选用前者。M3分布主要用于拟合交通量大于750veh／h且车队现象比较明显的交通流，根据线性回归得到的自由车比例α与交通量的函数关系，规定α应小于0．9。改进的M3分布适用性很广，但参数取值比较困难，可作为其他分布函数的有益补充。Thirty groups of expressway traffic data are used to verify the common distribution function of headways, showing that they are effective but they have significant overlaps among their applicability. To make their applicability clear, the characteristics of the distribution function are analyzed, and the selection strategies are put forward. The result shows that （1） negative exponential distribution can reflect the randomness of arriving of vehicles well and it is easy to use, so it is preferred when traffic volume is less than 250 veh/h per lane; （2） shifted negative exponential distribution with reasonable value of the minimum headways reduced the fitting error, while the order of Erlang distribution must be rounded, which enlarged the error, so the former is preferred when traffic volume is located at 250 - 750 veh/h; （3） M3 distribution is mainly used to fit the traffic flow when traffic volume is more than 750 veh/h and vehicle fleet are obvious, so the proportion of free cars named is limited to less than O. 9 according to the function between and traffic volume got by linear regression; （4） the improved M3 distribution has wide applicability, but the selection of parameters is difficult, so it is regarded as the useful backing function for other distribution functions.

关 键 词：交通工程 车头时距 数理统计 分布函数 高速公路 M3分布 

分 类 号：U491.114[交通运输工程—交通运输规划与管理]