Reasearch On Data Delivery In Vehicular Network Based On The Analysis Of Mobility Characteristics

Vehicular network,which applies the information science to transporta-tion field,is an important supporting technology for the ITS(Intelligent Transportation Systems).It is of great significance for creating a new service model of automobile and transportation,promoting the development and application of autonomous driving technologies,and improving traffic efficiency and travel safety.The mobility characteristics in vehicular network present unique features and regular patterns,due to the high-speed mobility,being limited by road topology,and being affected by traffic rules and driving behaviors.How to portray the mobility characteristics of the vehicles objectively and comprehensively is not only the important basis for data dissemination and routing,but also the main factor affecting the algorithm/protocol performance directly.Therefore,the primary problem to be solved is qualitative and quantitative analysis of the mobility characteristics of vehicles.In addition,the link duration between vehicles is very short.This incurs lack of stable end-to-end communication path and highly dynamic of network topology.These are main factors that affact the efficiency of instantaneous transmission in vehicular network.In order to improve the efficiency of data dissemination,vehicles can forward data through opportunistic communication,which expands the application of vehicular network and improve the efficiency of data dissemination.This dissertation first exploits the unique features and regular patterns of vehicles through some metrics in various aspects that reflects the power of data dissemination and the mobility characteristics of vehicular network.Secondly,an algorithm named VRank is proposed to identify some initial source nodes to offload traffic through opportunistic communication between vehicles.Finally,we analyzed the delay of safety messages propogation under the platoon-based driving pattern.The main research contents and innovations are as follows:(1)Due to the lack of real-world vehicular mobility dataset,several works have analyzed the mobility characteristics of vehicular network based on synthetic vehicular dataset,which cannot reflect the real vehicles behavior.In addition,these studies usually modeled vehicular network topology as a series of snapshots and neglected to take into account the temporal order between links.To overcome these shortcomings,we proposed a time-expanded graph to model the vehicular network and defined some metrics to portray the mobility characteristics in various aspects.In fact,time-ordered paths between two nodes can also be considered as opporuntistic communication routes over time.The temporal distance of a time-ordered path is corresponding to the delay of this route.Then,influence set,source set and reachability ratio are defined based on time-ordered path to reflect the power of opportunistic communication.The network diameter,centrality and connectivity are to describe the compactness and connectivity of vehicular network.Last,we analyzed the statistics distribution of these metrics to exploit the mobility characteristics of vehicular network on Beijing Taxi dataset.(2)Based on the analysis of mobility characteristics in vehicular network,we further exploited how to ease the burden of cellular network through opportunistic communication between vehicles.The offloading process usually contains two stages:initial sources selection and opportunistic data forwarding.For the issue of initial sources selection,we fir-st constructed the weighted reachability graph during a period of time based on time-expanded graph and time-ordered paths,which can capture the vehicles' transmission opportunities through instantaneous communication and opportunistic communication.Then,an algorithm named VRank was proposed to identify some influential vehicles to server as initial sources,which lead to spreading of content objects quickly and widely.As for content transmission,we should not only focus on the performance of delay,throughput,and packet loss rate but also pay attention to the user satisfaction in the process of data propagation.After that,we formulated the opportunistic data forwarding as a global utility maximization problem,which takes into heterogenous users' interests and future utility consideration.We proposed an optimal solution to solve that problem.Last,the effectiveness of our proposed approach was verified through extensive simulation using Beijing taxi dataset.The results showed that our proposed approach could spread the information to the whole network quickly and widely and satisfy users'interests.(3)Safety messages propagation via instantaneous transmission and opportunistic communication is the main task for vehicular network in order to improve the safety of roads and passengers.Mathematical modeling of the delay of safety messages propagation is extremely beneficial and complex due to the hi.gh dynamics of vehicles.Most previous works usually assumed that vehicles move independently and the interaction between vehicles is not taken into consideration.In the dissertation,we proposed an analytical model to describle the delay of safety messages propagation under platoon-based driving pattern.The statistics of time headway follows a lognormal distribution.In order to keep the safety distance between the following vehicle and preceding vehicle,we adopt the intelligent driver model to describe the mobility of vehicles.We first derived two important metrics:intraplatoon headway and interplatoon headway.Then,we analyzed the delay of safety messages propagation via instantaneous transmission between two vehicles in the same platoon and opportunistic communication between two successive platoons.Vehicles can also forward the safety message to an RSU or other vehicles traveling on the opposite direction to reduce the propagation delay.The best and worst scenarios are evaluated independently.The analytical model also takes into account different transmission situations and various system parameters,such as communication range,traffic flow,and platoon size.The effectiveness of the analytical model is verified through simulation and the impacts of different parameters on the transmission delay are investigated.