中文摘要：

城市交通系统是一个有人参与的、动态的和开放的复杂系统。传统的均值、方差等统计方法已难以直观地表达和发现城市交通流的潜在特征。而目前基于确定的交通流数学模型虽然能定量地揭示其非线性动力学特征，但存在着表达式复杂却精确度不够且灵活性欠缺和通用性不强的问题。本文探讨了城市行车速度的探索性分析方法的可行性，并以北京市四环内路段行车速度为例进行试验，结果表明，浮动车行车速度的探索性分析方法是有效的和实用的，可为交通流的特征分析特别是实时交通异常检测的研究提供参考思路。

英文摘要：

Urban transport system is a participatory, dynamic and open complex system. Conventional statistical methods such as mean and variance can no longer describe and discover potential characteristics of urban traffic flow. Furthermore, although existing mathematical models have been presented to reveal quantitatively nonlinear dynamics and chaos mechanism of traffic flows, it is still unavailable in practice because of its complicated expressions and lack of accuracy and flexibility. In this paper, exploratory analysis of travel speed based on the floating car data was discussed, and examined using real data sets of the fifth ring roadway sections in Bei- jing. This paper presented an exploratory analysis method for traffic anomaly detection based on quantile that conducts strong anti-dis- turbance and robustness performance. The results of applications showed that the proposed idea was feasible and effective, and could provide a referencial perspective on studying characteristics of traffic flow, especially traffic anomaly detection.