Research On The Key Technologies Of Data Delivery And Structure Optimization In Vehicular Networks

Vehicular networks are a class of open network systems with vehicles as core nodes.Equipped with onboard short-and middle-distance wireless communication devices,vehicles as well as vehicles and roadside facilities are interconnected so that information can be sensed and delivered,which will increase the safety and efficiency,and provide multiple information services to the vast number of drivers and passengers.Being a new type of wireless mobile ad hoc network,on the basis of onboard sensing and computing devices,vehicular networks become a kind of mobile sensing and information network as well,which has strong sensing ability and wide coverage.As an important part of the Internet of Things,vehicular network will be able to play a significant role in many fields of information applications in economic and social lives,including intelligent transportation and smart cities.For vehicular networks,it is an important task to support the circulation of information,to which data delivery is crucial in network systems.From the perspective of application,this thesis concentrates on issues of data delivery and structure optimization in vehicular networks,and carries out empirical researches on the basis of large-scale real-world data sets.We study on data delivery by exploring the topics of mobile information dissemination service,urban sensing data gathering service,and network structure optimization for data delivery in vehicular networks,and come up with optimized designs and schemes.The main contributions of this thesis are as follows:· In the mobile information dissemination service,we perform empirical study on the basis of large-scale real vehicular traces,and propose an innovative scheme for addressing the problem.The fast-varying topology of vehicular networks along time makes it hard to know the exact future information about the network.And to widen the range of spread of information as great as possible while the contents of them are still valid,especially picking out those best seed vehicles under a limited budget such that the information coverage in the whole network is maximal within a given time period is also of great challenge.To tackle the challenges,we define the problem of mobile information dissemination in vehicular networks first,and prove that it is an NP-hard problem.Afterwards,we carry out detailed analyses of the data sets of several vehicular networks,and have the observations that vehicular networks demonstrate clear sociality.The vehicular sociality is highly dynamic,while such dynamics also show strong temporal patterns and correlations.On this basis,we capture the evolution patterns of vehicular centralities,infer future node behavior and contact information in the whole network,which is then utilized to select the best set of seed nodes.We also take into account structural equivalence between vehicle nodes in the network to further improve the performance of information dissemination.Extensive simulations based on real vehicular traces have been conducted,and the results verify the efficacy and better performance of our proposed scheme,especially in those scenarios where the dynamics of vehicular sociality are more significant.· In the urban sensing data gathering service,we propose to perform unified data gathering service using urban vehicular networks as the basis.Data gathering is a core service developed to support users of various smart mobile devices in contributing sensing data generated by their devices to the computing center as the information support for making urban management decisions.In the proposed design,for their wide distribution and high activeness in the cities,public vehicles are used to accept data from surrounding users and hand over to computing center through wireless base stations deployed in the city.Meanwhile,several among the vehicles are hired as relays,which assist gathering from others with multicopy and multihop forwarding towards the base stations.However,both the limited budget shared by deploying base stations and hiring relays,and the great uncertainty about collection opportunities of candidate locations for base station deployment and vehicles to hire in future gathering processes put forward great challenges to the design of base station deployment and relay-based forwarding scheme.We propose an empirical approach for unified data gathering.To tackle the challenges,we first characterize collection performance as function of temporal data paths towards each candidate,and formulate the problem as a multiobjective optimization problem,which is proved to be NP-hard.On that basis,explicitly exploiting the sociality within hybrid data gathering networks,we reveal regular relations between the candidates and estimate expected importance of them in the future gathering processes with large set of real vehicular traces,based on which we develop a common algorithmic framework for deciding base station deployment and corresponding forwarding strategy.The efficacy of our proposed approach is verified via extensive simulations based on real vehicular data sets.· For more efficient data delivery in vehicular networks,especially the applications that allow the great number of passengers to share their information with each other,in this thesis,we investigate the crucial problem of time-constrained data delivery in vehicular networks,and propose a constructive approach to the optimization of network structures.However,the unique characteristics of vehicular networks put forward great challenge to the problem.There are no alwaysconnected forwarding routes between vehicles,merely relying on contacts of vehicles cannot guarantee that there are forwarding routes for data delivery formed within the given period of time.Thus,support of mobile communication networks has to be introduced.However,there is an intrinsic tradeoff between communication cost and delivery quality.Furthermore,there is great uncertainty about vehicular mobilities.To tackle the challenges,in our proposed approach,contact-based data forwarding and mobile communication connections are integrated for network connectivity enhancement.Our approach first exploits the planned trajectories of vehicles for predicting the expected inter-vehicle contact events and corresponding times,and derives the time-stamped future topology information of the network.Based on that,we design accordingly the algorithm for computing the configurations for the optimization of the network,which enhance the connectivity of the network while introducing minimum number of communication connections through the mobile network.Extensive simulations based on real traces collected from thousands of public service vehicles in two metropolises have been conducted,and results demonstrate the efficacy and better performance of our proposed approach.