基于交通大数据的城市道路交通状态短时预测研究

项目名称： 基于交通大数据的城市道路交通状态短时预测研究

项目编号： No.61304188

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 自动化技术、计算机技术

项目作者： 单振宇

作者单位： 杭州师范大学

项目金额： 24万元

中文摘要： 道路交通状态预测是智能交通系统的核心技术之一,有助于提高道路交通运行效率,缓解交通拥堵这一当前亟待解决的经济社会问题。本课题拟采用大数据分析的思路，解决道路交通状态预测方法在城市路网应用中存在的鲁棒性差、覆盖范围小、实时性低等问题，主要研究内容包括：设计多源异类海量交通数据的表示方法，提高存储和计算的效率；建模全数据模式下的交通状态变化规律，揭示路网交通状态的时空域关系；研究数据复杂性问题，设计建模数据的有效抽取方法；基于以上算法和模型的实现，构建一个高效、准确、覆盖范围广的城市路网交通状态预测平台，并通过理论分析和真实路网中的实验评估上述模型和算法的性能和可靠性。

中文关键词： 交通状态预测；交通大数据；智能交通系统；数据表示；时空域模型

英文摘要： As a core technology in intelligent transportation system, high prediction accuracy of road traffic condition can improve road traffic efficiency, ease traffic congestion - the urgent economic and social problems. This project intends to use big data analysis to solve the problem caused by road traffic condition prediction method applied in urban road network such as poor robustness, small coverage, and there exists lags between the real situation and the prediction. The main research scope includes the following aspects: designing representation method of multi-source heterogeneous mass transportation data to improve memory and calculation efficiency; modeling traffic state changing rule with whole data to reveal the spatial domain relationship under network traffic status; studying data complexity and designing effective extraction of modeling data. Based on the above algorithm and implementation of the model, a highly efficient, accurate urban road network traffic condition prediction platform covering a wide range will be built. And their performance and reliability will be evaluated by theoretical analysis and experimental evaluation of the real road network.

英文关键词： Traffic forecasting；Traffic big data；ITS；Data representation；Time-space domain model