基于时空自相关的道路交通状态聚类方法

城市道路交通状态的识别对交通管理部门进行交通管理控制、出行诱导，以及 道路设施改造具有重要意义.本文运用时空Moran 散点图探索城市道路交通的时空关联 性，并据此构建一种基于时空自相关预分类的道路交通状态层次聚类方法.运用本文所提 出的聚类算法，以北京市二环快速路外环方向的路段为例，进行聚类研究，并分析了各类 型路段的交通状态时空特性.案例研究表明，所提出聚类算法能对道路交通状态进行有效 判断，充分反映交通需求与路网结构之间的内在匹配关系.特别是畅通异质和拥堵异质两 种交通状态的提出，为识别高峰时段路网中的瓶颈路段和能力富余路段提供了一种新的 思路和方法，进而可为完善路网、缓解拥堵及制定交通管理措施提供依据.

Urban Traffic Status Clustering Method Based on Spatio-temporal Autocorrelation

Traffic condition identification through data mining is a crucial issue for Advanced Traffic Management Systems (ATMS). Spatio- temporal Moran scatterplot is used to study traffic status of urban road network in this paper. And accordingly, a hierarchical clustering algorithm considering presort of traffic status based on spatio-temporal autocorrelation is constructed. Finally, in order to demonstrate the feasibility and effectiveness of proposed hierarchical clustering algorithm, the clustering and classification for roads of the Second Ring freeway of Beijing are conducted. The results show that the proposed method can effectively reveal the spatio- temporal characteristics of different classes of roads and the relation of traffic demand to road network facilities. Especially, the introduction of heterogeneous congested and uncongested traffic in this paper makes it convenient and effective to recognize road sections with traffic dredging or bottleneck effect in the road network, which can provide foundations for infrastructure reform, congestion alleviating and traffic management measures formulating.