移动分布感知的车载自组网络数据分发

Mobile Distribution-Aware Data Dissemination for Vehicular Ad Hoc Networks

提出了一种面向发布/订阅系统基于车辆移动分布感知的事件分发策略MDA(mobile distribution-aware data dissemination).基于车流的自组织性及自稳性的特点,建立VANET(vehicular ad hoc network)下的发布/订阅模型,通过计算车辆与移动订阅者的相遇概率,预测订阅者的移动分布,并以此为依据实时部署和调度广播令牌在网络中的转发,从而有效地控制事件代理的分布,保证了数据传递的有效性.与已有相关研究相比,MDA采用的启发式算法,能够使事件代理的分布更好地适应网络环境的动态变化.此外,MDA采用了一种基于概率预测密度的令牌控制算法,通过实时地调整令牌的数量,进而控制事件代理的数量,降低了整个网络的负载.模拟实验结果表明,与现有的几种消息分发算法相比,MDA能以较低的网络负载和传输延迟获得较高的数据传输成功率.

This paper proposes MDA, a mobile distribution-aware data dissemination, based on Publish/Subscribe, for Vehicular Ad Hoc Networks （VANETs）. To take advantage of VANETs＇ self-organization and self-stabilization, this paper first establishes an appropriate Publish/Subscribe model in VANETs. Next, the subscribers＇ distribution is predicted by calculating the delivery probabilities between vehicles and subscribers. Third, based on the predicted distribution, adjust notification token＇s deployment and forwarding in the VANETs are used to achieve effective distribution of notification brokers （notification-token holder）. Compared with existing solutions, a novel heuristic algorithm is applied to alter the notification broker distribution and to adapt to a real-time VANETs situation. Furthermore, MDA can also reduce an overall network load of VANET by controlling the occurrences in broadcasting. Simulation results based on a real city map and realistic traffic situations show that the MDA performs much better in terms of delivery and delay ratios and have a higher network load than other solutions.