为了寻求合适的概率分布模型,准确描述间断流环境下路段人均行驶速度的变化规律,提出基于多源检测数据的人均行驶速度估算方法.选取正态分布及对数正态分布为基本形式,构建4种双峰概率分布函数.郑州市实际数据的拟合及检验结果表明,人均行驶速度的概率分布多接近N_N双峰分布及N_Log N双峰分布.考虑公交及社会车辆混行状态,通过模型对样本数据的适应性对比分析,结果表明,人均行驶速度概率分布规律在有公交专用道路段适合N_N双峰分布,在无公交专用道路段适合N_log N双峰分布.
In order to search appropriate fitting models for describing change rule of per capita travel speed in features of interrupted flow, an estimating method of per capita travel speed based on multi-source is provided. Normal distribution and lognormal distribution are chosen as basic forms to build four kinds of bimodal distributions. Based on the fitting result of real data in Zhengzhou, China, the feasibility of bimodal distributions is analyzed quantitatively by the goodness of fitting test and the adaptation to variance of sample data. Considered mixed traffic flow on bus and car, results indicate that N_N bimodal distribution is fitted for the road sections with transit lane better, and N_Log N bimodal distribution is fitted for the road sections without transit lane better.