最熟悉的陌生人：一种基于相遇模式挖掘与非确定性规划的车联网数据传递方法

项目名称： 最熟悉的陌生人：一种基于相遇模式挖掘与非确定性规划的车联网数据传递方法

项目编号： No.61300178

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： 礼欣

作者单位： 北京理工大学

项目金额： 23万元

中文摘要： 车联网市场应用前景广阔，蕴含着巨大的商机和研究价值，其中车联网中的数据传递机制是众多应用的核心技术。由于城市道路的限制和人类行为模式的影响，基于车辆行驶轨迹预测的车联网数据传递方法被普遍认为具有良好的前景。但传统的轨迹预测方法倾向于对个体轨迹的独立分析和判断，忽略了相遇模式背后潜在的人类行为共性和社会属性的相似性对行驶轨迹的影响。本项目充分分析了车联网数据传递算法的研究现状，依据车联网真实数据提出了一种面向相遇模式挖掘的、与时间相关的轨迹预测模型；以预测的轨迹、数据投递率、时间延迟为参数和优化条件，提出一种基于非确定性序贯决策模型的车联网数据传递新方法；并通过搭建仿真验证平台，对提出的模型和算法进行分析和评价。

中文关键词： 车联网；数据传递；路侧单元；轨迹；地点预测

英文摘要： Vehicular networking is considered to be a very promising and prosperous market, which contains many business opportunities and is of great research value. For most vehicular applications, the data delivery mechanism is the key technology. In literature, the vehicular mobility prediction has been adopted for the data dissemination by considering the road layout and the human behavior simultaneously. However, the conventional mobility prediction focused on the individual tracks and is lack of the analysis of the human behavior patterns and the similarities of the social attributes which lead to the "occasionally" encounters. After conducting a thorough investigation of vehicular data dissemination algorithms, we propose a time dependent encountering analysis-based mobility prediction method, where the graph model is used to exploit the mobility over the real historical data. Then we propose a novel data dissemination method based on sequential decision making algorithom which optimize the problem to obtain the best action in terms of the data delivery rate and the average delay by considering the exploited mobility (the proper node for the next hop) for the data delivery. We will also develop the simulation system to verify the proposed methods.

英文关键词： Vehicular Networks；Data Dilivery；Roadside Units；Trajectory；Location Prediction