项目名称: 面向交通流微观结构的时空特征子空间分析方法研究
项目编号: No.41501442
项目类型: 青年科学基金项目
立项/批准年度: 2016
项目学科: 天文学、地球科学
项目作者: 赵玲
作者单位: 中南大学
项目金额: 20万元
中文摘要: 随着城市交通拥堵现象的日益严重,研究影响宏观交通流现象形成及变化的本质因素,发现交通流的运行与演化规律,对城市交通流管理具有重要意义。本项目在融合分析多种交通流数据(如浮动车数据、手机定位数据、固定检测器、社交媒体数据等)的基础上,力图突破传统的基于模型或数据驱动的交通流解析方法,采用数理方法刻画人的出行决策、城市功能分区和交通流的网络分布,运用稀疏表达等手段从宏观交通流现象中提取内在的微观结构特征,即“特征子空间”。进而,利用时空数据分析技术发现特征子空间的时空分布和社区关联特征,进一步揭示网络交通流宏观性质如何通过微观局部的相互作用涌现机理,为解析网络交通流的运行与演化特性、精确分析和预测交通流、实现交通流的有效管控等提供理论基础。
中文关键词: 时空特征分析;交通流;时空聚类;稀疏表达;特征子空间
英文摘要: With the worsening traffic congestion, it is very important for traffic flow management to research the internal-factors of the macroscopic traffic flow and reveal the rule of urban traffic flow. The project breaks through the traditional analytical methods based on model or data-driven. Various traffic flow data, such as probe cars, mobile phone location, fixed detectors and social media etc, are fusion analyzed. The passengers’ trip decision and path selection and the network distribution of traffic flow are depicted by mathematical methods. The internal micro-features of the macroscopic traffic flow, which name is feature subspace are extracted by sparse representation. Then, the spatio-temporal distribution of feature subspace and the community associated features are discovered with spatio-temporal analysis. So as to further reveal the mechanism that macro-properties of network traffic flow emerge with the micro-factors. All of these works provide the theoretical basis for interpreting characteristic of network traffic flow, precisely analyzing and predicting traffic flow and realizing the effective control of traffic flow.
英文关键词: Spatio-temporal feature analysis;Traffic flow;Spatio-temporal clustering; Sparse representation;Feature subspace