节点文献

城市场景车载Ad Hoc网络单播路由机制的研究

Research on Unicast Routing Mechanisms for Vehicular Ad Hoc Networks in City Scenarios

【作者】 李元振

【导师】 廖建新;

【作者基本信息】 北京邮电大学 , 计算机科学与技术, 2010, 博士

【摘要】 近年来,车载通信成为国内外研究的热点。实现汽车间的无线通信可以给人们的生活带来极大的便利。例如,在公路上行驶时,人们可以通过通信系统来了解路段的情况,根据是否发生事故或堵车来及时调整路线。同时交通系统也可以根据所收集到的车流量和路况信息,及时调整红绿灯时间,使交通系统达到最高效率。从上述可以看出,引入车辆间无线通信技术可以极大地提高行车安全以及改善交通流量状况。在车辆通信方面,移动自组网技术具有很多有利条件,例如在低延时、网络健壮性以及网络拓扑结构等多个方面具有更大的优势。车载Ad Hoc网络(VANET:Vehicular Ad Hoc Networks)是在配备无线网络接口设备的汽车之间形成的一种特殊的无中心移动自组织网络,是目前MANET (MANET:Mobile Ad Hoc Networks)最具前途的应用领域之一。由于车载Ad Hoc网络中所有节点地位平等、自组织、无需事先布置网络设施或中心控制节点,而且网络节点高速移动会导致网络拓扑动态变化等,作为车载Ad Hoc网络关键技术之一的路由算法在很大程度上决定了整个车载Ad Hoc网络的性能。因此,如何设计出有效的路由算法从而保障信息及时、可靠地传递,一直是车载Ad Hoc网络的研究热点和难点。论文总结了作者在城市场景下车载Ad Hoc网络中单播路由机制的主要研究工作,包括:提出了一种基于地理和交通信息的车载Ad hoc网络路由算法;提出了一种基于竞争转发的车载Ad hoc网络路由算法;论文对基于竞争转发的车载Ad hoc网络路由算法进行了分析,给出了保证抑制策略以一定概率执行成功的条件下,节点定时器参数的下限值;提出了稀疏车载Ad hoc网络中端到端时延计算公式,并提出了一种稀疏车载Ad hoc网络中时延参数计算评估算法DEA (Delay Evaluation Algorithm)用于衡量路由算法的时延性能的优劣;论文对一维车载Ad Hoc网络若干连接特性进行了分析;基于若干连接特性的分析,论文提出了一种城市场景下车载Ad Hoc网络中基于粒子群算法的服务节点部署方案,其中服务节点可以是节点辅助路由中的辅助节点,也可以是提供Internet接入的网关。对于上述协议和算法,均进行了仿真验证和性能分析。论文对研究过程中取得的主要创新工作进行了详细阐述。这些创新工作简要归纳如下:(一)基于地理和交通信息的路由算法。论文提出了一种基于地理和交通信息的车载Ad hoc网络路由算法,该算法利用了实时的道路交通信息和车辆移动位置预测进行数据转发,并采取暂存转发策略。该算法还能处理目的节点移动的情况。仿真结果表明该算法比类似算法具有更高的数据包递交率和更低的时延。(二)基于竞争转发的路由算法。论文提出了一种适用于城市场景的基于竞争转发的车载Ad hoc网络路由算法,摒弃了基于地理和交通信息的路由算法HELLO消息的使用,利用分布式的竞争策略根据邻居节点所处的地理位置信息进行下一跳节点的选择。数据转发使用锚点路由,同时利用了道路交通信息和网络拓扑结构信息进行锚点选择。采取暂存转发策略处理网络分割现象。仿真结果表明该算法比其他类似算法具有更高的数据包递交率和较低的数据延迟。论文对城市场景下的基于竞争转发的车载Ad Hoc网络路由算法进行了分析,给出了保证抑制策略以一定概率执行成功的条件下,节点定时器参数的下限值。仿真和分析结果的对比验证了分析的有效性。(三)稀疏车载Ad Hoc网络中时延评价算法。网络分割的情况下,节点使用暂存转发机制提高数据包的递交率。论文提出了稀疏车载Ad hoc网络中端到端时延计算公式,并提出了一种稀疏车载Ad hoc网络中时延参数计算评估算法DEA (Delay Evaluation Algorithm)用于衡量路由算法的时延性能优劣。算法以端到端时延最小化为目标,以节点运动过程中形成的相邻关系为输入,计算得出源节点到目的节点的具有最小时延的路径以及最小时延值。DEA同时能够应用在运动规律确定的车载Ad Hoc网络中进行路由决策。(四)一维车载Ad Hoc网络若干连接特性分析。论文对节点间距服从指数分布的车辆网络中的连接特性进行了分析,给出了(1)车辆i和车辆j之间的连接概率,(2)长度为L的区间上网络连接概率,(3)位于0点的车辆通过多跳通信覆盖地理点L的概率。蒙特卡罗实验验证了论文分析的有效性。论文的结果可以用于车载Ad Hoc网络中路由决策和节点部署。(五)基于连通性的车载Ad Hoc网络中服务节点部署方案研究。车载Ad Hoc网络中部署服务节点可以辅助路由或者提供Internet接入等服务。论文基于车载Ad Hoc网络中连接特性的讨论,提出了车载Ad Hoc网络中服务节点部署模型,并基于粒子群算法给出了一种城市场景下车载Ad Hoc网络中服务节点部署方案。通过仿真与实验,验证了算法的有效性。

【Abstract】 In recent years, there is a growing interest on the research and deployment of inter-vehicle communication system. Inter-vehcle communication can bring great convenience to people. For example, when traveling on the road, people can get the traffic information through the road communication system and adjust their travel line timely depending on the traffic information. Furthermore, transportation system can also adaptively adjust the traffic lights by traffic flow and traffic information to enhance efficiency in Intelligent Transportation Systems(ITS). Be seen from the above, the inter-vehicle wireless communications technology can greatly improve traffic safety and improve traffic flow conditions. In inter-vehicle communications, mobile ad hoc network technology has many advantages, such as low latency, network robustness and network topology.VANET(Vehicular Ad Hoc Networks) is a specific type and one of the most promising applications of mobile ad hoc networks, which is formed by vehicles that are equipped with wireless communication devices. As Vehicular Ad Hoc networks is self-organization without network infrastructure or central control node and high dynamic topology due to high speed, how to design an efficient routing protocols to delivery packets opportunely and reliablely has been a hot topic for Vehicular Ad Hoc Network.In this dissertation, the author’s major research work in this area is summarized as follows:A novel geography and traffic based routing algorithm for VANETs (Vehicular Ad hoc Networks) in city environments was proposed. A contention-based forwarding routing algorithm for VANETs in city environments was proposed. This dissertation proposed an analysis of the contention-based forwarding (CBF) routing algorithm for VANETs in city environments and the lower bound of the timer was given on the condition that the suppression scheme accomplished successfully at give probability. This dissertation gave the formula for End-to-End delay in spare Vehicular Ad hoc Networks, and an algorithm named DEA(Delay Evaluation Algorithm) was proposed to evaluate the End-to-End delay in routing algorithm. In this dissertation, the probabilities of Vehicular Ad hoc Network connectivity for difference cases were derived. Based on the probabilities, a service node placement strategy for Vehicular Ad hoc Networks(VANETs) in city scenarios was proposed. And the service node can be auxiliary nodes in routing algorithm or Internet Access Gateway. A simulation platform is constructed to verify the correctness and to analyze the performance of the above algorithms.(1) Geography and Traffic Based Routing for Vehicular Ad Hoc Networks in City Scenarios. A novel geography and traffic based routing algorithm for VANETs (Vehicular Ad hoc Networks) in city environments was proposed, which using real time traffic information and mobility prediction to forward data with the idea of "carry and forward". Furthermore, this protocol can work well where the destination node is moving. Simulation result shows that it achieves significant delivery rate improvement and lower delay compared other similar routing approaches.(2) A Contention-Based Forwarding Routing Algorithm. A contention-based forwarding routing algorithm for VANETs in city environments was proposed, which using distributed contention-based strategy to select the next hop based on the positions of the neighbor nodes without HELLO messages. The novel scheme forwards packets along an anchor-based path which is determined according real time traffic information and the topology of the network, with the idea of "carry and forward" to deal with partitioned networks. Simulation result shows that it achieves significant delivery rate improvement and lower delay compared other similar routing approaches. This dissertation proposed an analysis of the contention-based forwarding (CBF) routing algorithm for VANETs in city environments, which is very important to determine the value of the timer used in CBF. And the analysis is one beneficial and important component of CBF. The lower bound of the timer was given on the condition that the suppression scheme accomplished successfully at give probability. The analysis is validated by comparing simulation and analysis results. (3) A Benchmark for Delay Performance in Spare Vehicular Ad hoc Networks. In spare networks, node forwards packet using "carry and forward" to improve the delivery rate. This dissertation gave the formula for End-to-End delay in spare Vehicular Ad hoc Networks, and an algorithm named DEA(Delay Evaluation Algorithm) was proposed to evaluate the End-to-End delay in routing protocol. DEA aims to minimize End-to-End delay. Taking the neighbor relationship of nodes as input, DEA outputs the minimal End-to-End delay and the routing path with the minimum delay. DEA can also be used in a network routing where nodes move regularly.(4) On the Connectivity of One-Dimensional Vehicular Ad Hoc Networks. In this dissertation we analyzed connectivity of one-dimensional Vehicular Ad Hoc Networks where vehicle gap distribution can be approximated by an exponential distribution. The probabilities of Vehicular Ad hoc Network connectivity for difference cases are derived:(1) The probability of connectivity of the ith vehicle and the jth vehicle(j>i). (2) The probability of connectivity of the vehicles in the interval [0,L](L>0) (3) The probability of connectivity of a fixed node at point zero and the geographical point L(L>0). The analysis is validated by computing the probability of network connectivity and comparing it with the Mont Carlo simulation results.(5) Connectivity-based Service Node Placement in Vehicular Ad Hoc Networks. This dissertation proposed a service node placement strategy for Vehicular Ad hoc Networks in city scenarios. Firstly based on the probability of connectivity and the characteristic of VANETs, the model of optimal placement of the service nodes was described. Furthermore, one solution based on PSO(particle swarm optimization) for this optimal problem was given. Finally, we validated our scheme by simulation.

节点文献中: