车联网络作为物联网在智能交通系统中发挥着重要应用，是车辆间通信或者车辆与其它物体通信的一种载体。随着现代智能交通系统的快速发展，车联网络受到工业界和学术界越来越多的重视。由于车联网络具有其自身的一些特性，如，拓扑随时变化、行驶方向规律、延时受限等，车联网络性能受到严重限制。传统的一些无线网络技术并不能很好的适用于车联网络。为更好的分析车联网络特性、提高车联网络性能，本论文针对车联网络信道传输特性进行建模，并对车联网络应用中的一些关键技术进行展开研究，同时对不同应用场景进行建模和分析。本论文主要工作和创新点包括以下六个方面： 首先，论文研究车联网中车辆访问信道时信道切换引起的高数据包碰撞问题，建立了信道切换分析模型来分析和计算信道切换引起的数据包传输碰撞概率和传输延时。在碰撞概率数学模型和传输延时数学模型建立过程中，考虑到信道切换中车辆动态竞争行为，揭示了数据包突然到达时引起的高竞争访问现象。针对这一问题，本论文设计两种解决方案：等间隔划分信道和随机间隔划分信道，并且分别给出两种解决方案下车联网络传输数据包碰撞概率模型和传输延时模型。通过仿真验证可知，提出的数据包传输碰撞概率模型和传输延时模型能够描述信道切换引起高竞争访问现象的有效性和准确性。同时，仿真结果表明提出的等间隔划分信道方案和随机间隔划分信道方案能够有效减少信道切换时引起的高竞争现象，降低数据包传输碰撞概率和数据包传输延时，同时提高车联网络吞吐量。 其次，论文研究车载D2D中继网络中带有多个用户时传输容量问题。通过建立传输容量理论分析模型来分析和计算车联网络传输容量。在分析和计算过程中，假设D2D中继网络中多用户(服务车辆和辅助车辆)工作在正交模式。采用随机集合理论，服务车辆和辅助车辆在车联网络中分别服从泊松点分布。服务车辆与服务车辆通信时，计算辅助车辆存在概率和辅助车辆链路平均距离表达式从而获得传输成功概率。在辅助车辆中继传输扩大传输距离特点下，进一步获得辅助车辆辅助情况下服务车辆与服务车辆通信时传输容量，且考虑到车辆密度和传输功率因素的影响。除此之外，分析了车联网络传输容量与服务车辆与服务车辆之间通信变化的直接链路距离的关系表达式。仿真结果表明通过辅助车辆且调整通信服务车辆之间通信距离、车辆密度等因素可以有效改善传输容量。 再次，论文研究5G下车联网络中可靠链路传输问题，建立一个链路服务质量指示器模型来分析和计算传输链路的传输可靠性。在毫米波车联网络中，通过链路传输成功概率和传输延时确定链路服务质量，其中不同的传输成功概率和传输延时相应权值反映链路不同的服务请求。在链路服务质量指示器模型建立过程中，分析和对比了链路服务质量指示器与车辆数目和不同车辆传输距离之间的关系。在链路服务质量指示器模型基础上，提出一个基于车载云平台的最优链路选择算法。通过实验仿真验证一定程度上增加车辆数目能够有效提高车联网络链路服务质量，同时评估车辆数目、车辆传输距离及不同权值因子对于车辆网络链路服务质量的影响。对于不同服务需求，车联网络中总是存在一条最优链路满足用户服务需求。 接着，论文研究两跳中继车联网络传输链路中断性能问题，分析直接通信链路和中继通信链路两种情况，主要建立中继链路中断概率数学表达式模型分析和计算中继传输链路中断概率。在中继链路传输链路中断概率模型中，假设传输信道为瑞利衰落信道，通过计算中继辅助链路传输信噪比得到车辆与车辆通信的概率密度函数，进一步获得车辆与车辆之间通信链路的中断概率，从而反映两跳链路传输信噪比和中继增益带来的影响，其中链路中断与否的判决依据链路传输容量与设定阈值之间的对比。另外，设计一种车载设备功能结构以实现提出的最优中继路径选择算法。仿真结果验证了中断概率模型的有效性和准确性。对于不同的链路信噪比，仿真结果表明中断概率与链路信噪比之间的关系。 此后，论文研究基于竞争车联网络中不同车辆负载情况下车辆访问信道时退避算法。提出一种动态自适应成功-碰撞退避算法改善基于竞争的车联网络性能。由于碰撞概率反映了车联网络性能且受到网络中车辆负载及竞争窗口变化的限制。阈值设定反映了不同车辆密度负载。另外，提出的算法通过数据包传输成功次数或碰撞次数相应的减小或者增加窗口大小以优化车联网络吞吐量和传输延时。通过仿真结果可以验证提出的动态自适应成功-碰撞退避算法能够有效改善车联网络吞吐量和传输延时方面性能，且给出了竞争窗口、车辆密度等对于车联网络性能影响，为车联网络在实际场景中合理部署提供理论参考依据。 最后，论文研究了车联网络中双天线车辆和单天线车辆共存情况下访问信道的公平性问题。建立了一个公平性模型来分析和计算访问信道的公平性。在模型的建立过程中，依据双天线车辆和单天线车辆访问信道时，信道访问持续时间不同，影响不同车辆的吞吐量，从而造成不公平性问题。以车辆访问信道时吞吐量为目标，分别分析和计算双天线车辆和单天线车辆在车联网络中访问信道时的吞吐量，将单个双天线车辆和单个单天线车辆吞吐量相等判定为访问信道公平。通过仿真实验可以验证提出的公平性访问信道模型的有效性和准确性。结果表明双天线车辆与单天线车辆访问信道的公平性与网络中双天线车辆数目、单天线车辆数目、双天线车辆最小竞争窗口和单天线车辆最小竞争窗口之间的关系且总是存在合适的竞争窗口大小保证访问信道的公平性。 关键词：车联网络、信道建模、D2D、5G、链路选择、公平性

 Vehicular networks is a carrier for communication between vehicles or for vehicle to communicate with other things and plays an important role in intelligent transportation system(ITS). With the rapid development of modern ITS, vehicular networks have received increasing attention from the industry and academia. As of vehicular networks has its own characteristics, such as changes in the topology, regular driving directions, and limited delay, the performance of the vehicular networks is severely limited. Some traditional wireless network technologies are not very suitable for vehicular networks. In order to analyze characteristics of vehicular networks and improve the performance of vehicular networks,this dissertation focuses on the modeling of channel transmission characteristics of vehicular networks and studies some key technologies on vehicular networks applications. Furthermore, this dissertation focuses on the modeling and analysis of different applications in different network scenarios. The main work and contributions of this dissertation including the following six aspects: Firstly, high channel congestion phenomenon following a channel switching when vehicle access channel in vehicular is studied and a theoretical analysis model with channel switching is established for analyzing and calculating the packet collision probability and transmission delay caused by channel switching. In process of establishing the model, vehicle dynamic contention behavior reveals high channel congestion phenomenon as of burst packet arrival On basis of model, two solutions including a centralized equal-spaced algorithm and a distributed random-spaced algorithm are proposed and packet transmission collision probability and transmission delay are also proposed respectively for two solutions. The accuracy and effectiveness of the packet transmission collision probability model and transmission model which depict high channel congestion phenomenon following a channel switching are verified through simulation results. Meanwhile our proposed solutions can effectively mitigate high channel congestion phenomenon caused by channel switching, reduce packet transmission collision probability and transmission delay and improve system throughput. Secondly, transmission capacity in vehicular D2D relay networks with multi-user case is studied and a transmission capacity theoretical analysis model is established for analyzing and calculating the transmission capacity. In the process of analyzing and calculating, we assume multi-user including service vehicle (SV) and helper vehicle (HV) in D2D relay network work in overlay mode. According to stochastic geometry theory, SVs and HVs all are modelled as Poisson Point Process (PPP) in vehicular networks, respectively. When SV communicates with SV, the HVs existence probability and the expectation distance with HV link are achieved to calculate the successful transmission probabilities. In the characteristics of enlarging transmission distance for HV, transmission capacity between SV and SV with assisted HV is further achieved and reveals the influence of vehicles density and transmission power. Besides, the relationship between transmission capacities and variable direct link distance of SV to SV communication are also analyzed. Simulation results indicate that transmission capacity can be improved by a HV and influenced by some system parameters, including communication distance between SV and SV assisted by HV, vehicle density and so on. Thirdly, reliable link transmission in 5G enabled vehicular network is studied and a link quality of service (QoS) indicator (LQSI) model is established for analyzing and calculating the reliability of transmission link. In millimeter wave vehicular network, we consider link transmission success fill probability and transmission delay to judge link QoS and different weight lactor for transmission success fill probability and transmission delay reflect different link QoS, respectively. In establishing the LQSI model, the relationship between LQSI and vehicles number and transmission of vehicle is analyzed and compared. Based on the proposed LQSI model, an optimal selection algorithm based on vehicular cloud (VC) is proposed. Simulation results indicate that link QoS of vehicular network can be improved by increasing number of vehicles. Simultaneously, the influence of vehicles number, transmission distance and different weight factors for link QoS is evaluated in vehicular. For different service requirements, there is always an optimal link to satisfy user service requirements. Fourthly, the outage performance of two-hop vehicular networks, including direct communication and indirect communication is studied and an outage probability egression model is established for analyzing and calculating outage probability of relay transmission link. In establishing outage probability model, the probability density function for vehicle to vehicle (V2V) communication is obtained by calculating the over transmission signal-noise-ratio (SNR) of the relay-assisted link under Rayleigh fading channels. Then, the outage probability of V2V communication link is further achieved and reflects the influence of transmission SNR for two-hop links and relay gain. The ratio of transmission and threshold judge the link is good or bad. In addition, an optimal relay selection algorithm for vehicle device function structure is designed. The accuracy and effectiveness of outage probability model is verified through simulation results. For different link SNR, simulation results indicate the relationship of outage probability and link SNR. Then, the backoff algorithm under different traffic load conditions in contention-based vehicular network is studied and a dynamic adaptive success-collision (DASC) backoff algorithm is proposed to improve the performance of vehicular networks. The collision probability reflects the performance of vehicular networks and restricted by vehicles traffic loads and the variation of contention window (CW). The threshold is defined to reflect different traffic loads. In addition, the proposed algorithm improves throughput and transmission delay of vehicular networks by reducing CW value after consecutive successful transmission times and increasing CW value after consecutive collision times. Simulation results indicate that the proposed DASC backoff algorithm could improve the throughput and transmission delay of vehicular networks and give out the influence of CW and vehicles density in vehicular networks which can provide theoretical reference for reasonable deployment in practical environment. Finally, the feirness problem of the access channel under the coexistence of dual-radio vehicles and single-radio vehicles in vehicular networks is studied and a fairness model is established for analyzing and calculating the fairness of the access channel In the process of establishing the model, the difference duration for dual-radio vehicles and single-radio vehicles cause the different throughput and further to causing unfeirness problem. With the goal of throughput when the vehicle accesses the channel, we analyze the throughput of dual-radio vehicles and single-radio vehicles,respectively. The access channel is feirness if one dual-radio vehicle and one single-radio vehicle have same throughput. The accuracy and effectiveness of the fairness access channel model are verified in simulation results. Simulation results indicate that the relationship between the feirness of the access channel of a dual-radio vehicle and a single-radio vehicle and the number of dual-radio vehicles, the number of single-radio vehicles, the minimum contention window of a dual-radio vehicle,and the minimum contention window of a single-radio vehicle. In addition, there is always an appropriate CW to ensure fairness access channel. Key Words: vehicular networks, multi-channel modeling, D2D,5G,link selection, fairness