Research On Network Selection Scheme Of Heterogeneous Vehicle Network Based On Intuitionistic Fuzzy Decision Making

Internet of vehicles(IoV)is a wireless communication and information exchange system based on intra-vehicle network,inter-vehicle network and vehicular mobile internet,through people,vehicles,roads,Internet collaboration to provide users with personalized information services.In urban environments,there are access networks based on different wireless technologies,such as cellular networks,WiFi networks,and automotive ad hoc networks.These network target scenarios and target users are different,and the technical indicators such as transmission rate,coverage,and delay are different.How to design a network selection scheme for heterogeneous IoV environment to ensure the quality of service(QoS)acquired by users has become a research hotspot and a difficulty for car networking,and is also the focus this thesis.Considering the network selection problem of IoV,this paper considers the rapid change of signal strength caused by high-speed vehicle movement.Based on the analysis of relevant protocol standards and the characteristics of heterogeneous IoV,this paper mainly studies the network selection algorithm and has achieved the following achievements :First,the architecture of the system with switch server as the core is given.Then,the overall mechanism based on IEEE802.21 vertical handover protocol is designed for network information acquisition,decision-making and handover execution.Based on this,in order to reduce the ping-pong effect caused by the switching too frequently between access networks,the decision-making information from this handover to the next handover is considered as a whole,the signal strength in this period is predicted,and the latency of the access network is acquired from the server.Then,an intuitionistic fuzzy evaluation matrix of candidate networks is generated.Using QoS as an index,the ranking of alternative networks is obtained by solving the intuitionistic fuzzy decision problem by using the Technique for Order Preference by Similarity to an Ideal Solution(TOPSIS).Through the simulations under the expressway scenario,it is verified that our mechanism can significantly reduce the ping-pong effect and improve the QoS compared with other mechanisms.Secondly,considering that different applications such as streaming media and online games have different requirements on network signal strength,delay and other parameters,the multiapplication network selection algorithm is studied.By selecting the utility function such as Sigmoid,exponent and so on to describe the difference of demand,the intuitionistic fuzzy evaluation matrix of each application to the alternative network is obtained by combining the signal strength and the network congestion degree parameter.The problem of network selection when vehicle users use multiple applications simultaneously is transformed into an intuitionistic fuzzy group decision making problem.By giving higher evaluation weights to similar evaluation matrices,more applications are compatible.Taking the positive and negative ideal conditions in the evaluation matrix as a benchmark,this intuitionistic fuzzy decision-making problem is solved and the purpose of meeting the multi-application QoS requirements is achieved.In order to further improve the utilization of RSU,a network recommendation mechanism is proposed to serve users without automatic network switching function solving the problem of adjusting the number of vehicles receiving recommendation information.At last,the simulation results of highway scenario prove the effectiveness of this algorithm in QoS enhancement and recommendation mechanism in multi-application scenarios.Finally,a high-precision signal strength sensing platform based on PXI spectrum analyzer is built,which can realize the joint acquisition of signal strength and GPS information of 18 network frequency bands.After that,using the platform to carry out the campus signal strength perception test,so as to build a signal strength database.Android application based on the MVP model which can adapt to most of the models of signal-strength is developed to reduce the cost of testing.After testing in selecting representative sections of campus,the LSTM method was used to predict the signal strength.Finally,the influence of the vehicles on the signal strength was analyzed,which laid a good foundation for studying the spatial and temporal variation of the signal strength in the urban area.To sum up,this thesis designs the whole mechanism for network selection in accordance with 802.21,and designs the network selection scheme of vehicle users in single application and multi-application scenarios based on intuitionistic fuzzy,so as to meet the QoS requirements of vehicle users.And developed a signal strength awareness platform based on PXI and Android.The research of this thesis helps to promote the development of IoV convergence technology and related applications.