面向车联网的交通网络涌现行为建模

Vehicular ad hoc netWorks(VANETs) industry is facing unprecedented opportunities and challenges, but it is difficult to meet the requirements of improving the efficiency of the transportation network system,as well as the need to guide public safety and convenient travel. The complex networks theory is applied to the field of VANET. The relationship between road network and VANET is researched from the perspective of complex network. Different modals of road network and VANET are modelled. Transportation network evolution process is clarified based on the multimodal road network and VANET, and the topological feature of the transportation network is researched. Emergency behavior model of transportation network is proposed for VANET and road network. The operation mechanism of the transportation network system is clarified through the model. The method of state analysis and trend evaluation for transportation network is created. Eventually the need for state analysis, forecasting and early warning, emergency command and etc of the transportation network is met.

车联网产业面临前所未有的机遇与挑战，但却难以满足提高交通网络系统的效率，以及引导公众安全便捷出行之需求。本项目将复杂网络理论引入车联网领域，从复杂网络视角研究路网与VANET之间的关系，对路网和VANET的多种模态进行建模，阐明基于路网和VANET多种模态的交通网络演化过程，进而研究交通网络的拓扑特性，提出面向路网和VANET的交通网络涌现行为模型，通过该模型阐明交通网络系统的运行机理，建立交通网络状态分析和态势评价方法，最终满足交通网络态势分析、预测预警、应急指挥等需求。

本项目从研究路网和车联网这两个复杂网络的相互关系问题入手，在分析交通网络系统中存在的多重模态的基础上，研究路网模态下的交通网络拓扑特性和涌现行为，构建了路网和车联网模态数学模型。建立了面向路网监测的车联网组网方法，在RSU部署数量受限的条件下，给出如何选择最优位置点覆盖更多车辆的RSU部署方案。提出了利用分布式时分多址协议未分配时隙的车联网协作通信和数据重传方法，显著提高车联网数据包发送成功率，有效降低丢包率和数据包传输时延。研究了城市道路短时交通流量预测问题，考虑多种数据源影响因素，抽象路网结构，结合影响因素的时空依赖特性，提出一种基于组合最小二乘支持向量机的城市交通网络短时流量预测模型。针对车联网动态变化的环境中如何合理使用传感器节点受限资源并获得较高应用性能的问题，将强化学习方法用于传感器节点的任务调度，提出了一种基于Q学习和规划的传感器节点任务调度算法。