5G异构网络干扰建模与仿真

发布时间：2018-03-28 03:11

本文选题：5G　切入点：超密集网络　出处：《西南交通大学》2017年硕士论文

【摘要】：为了满足数据流量井喷式增长以及用户体验速率提高10～100倍的需求,下一代移动通信(5G)异构网络部署将使用超密集网络(Ulltra-DenseNetwork,UDN)。超密集网络的部署是通过在原有的异构网络里增加大量的低功率基站节点(小小区Small Cell),包括微蜂窝基站(Microcell)、微微蜂窝基站(Picocell)、毫微微蜂窝基站(Femtocell,又称家庭基站)、端到端节点(D2D)等。5G系统中节点的部署密度将超过现在的10倍以上。虽然超密集网络缩小了终端用户与节点基站之间的距离,使得网络频谱效率大幅度提升,系统容量得到扩展,但是低功率节点数目的剧增,节点间距离的缩小,越来越密集的网络节点部署使得网络拓扑结构更加密集化、异构化和复杂化,这样就导致干扰环境更加复杂化。因此,研究5G异构网络干扰的建模具有重要意义。传统的正六边形网格模型较为固定但并不精确,对于未来多层异构化的5G超密集网络并不适用。对于这个问题,近几年比较常用的方法是采用随机几何里的泊松点过程来建模网络部署,且假设每一层的基站位置符合相同的点过程或者使用完全随机的泊松点过程。由于宏蜂窝小区边缘区域(盲区)、宏蜂窝小区热点区域(忙区)以及D2D通信(Device to Device Communication)等特殊区域中,基站的分布并不是一样的,且不同场景下用户终端的分布也不同,因此,现有简单的泊松点过程模型将不再适用。针对这个问题,本文针对不同的场景利用不同的点过程模型实现了不同的网络模型的模拟,并对其干扰进行了分析。首先,针对宏小区边缘区域建立了具有层间相关性(不同类型基站间的空间相关性)的双层 MPP-PHP(Matern Hard-core Point Processes-Poisson Hole Process)干扰模型。通过与传统正六边形模型和现有简单的MPP-PPP(Poisson Point Process)模型进行仿真对比,分析验证了本模型的准确性和适用性,并仿真分析了该模型的干扰分布和PHP排斥半径对系统性能的影响,得到了在本文给定仿真参数下使系统性能最好的最佳排斥半径。其次,针对宏小区热点区域建立了具有层内相关性(同一类型基站间的空间相关性)的双层MPP-MCP(Matern ClusterProcess)干扰模型。通过与传统正六边形模型和现有简单的PPP-MCP模型进行仿真对比分析,验证了本模型的准确性和适用性,并仿真分析了该模型的干扰分布。最后,针对D2D场景建立了具有层间相关性的三层MPP-PHP-PPP干扰模型。与前面所建的双层MPP-PHP模型进行仿真对比分析,验证了本模型的准确性,并仿真分析了该模型的干扰分布和D2D对间距离对系统性能的影响,得到了在本文给定仿真参数下D2D对之间的距离对系统性能有一定的影响,随着距离的增大系统性能会变差,但影响并不是很大的结论的结论。
[Abstract]:In order to meet the demand of data flow blowout growth and the increase of user experience rate by 10 to 100 times, Ulltra-DenseNetworkUDNs will be used for deployment of heterogeneous networks in the next generation of mobile communications. The deployment of ultra-dense networks is achieved through the addition of a large number of low-power base station nodes in the original heterogeneous networks (small cell Small cell, including microcellular base stations, microcellular base-stations, microcellular base-stations, microcellular base-stations, microcellular base-stations, microcellular base-stations, microcellular base-stations, and microcellular base-stations). The deployment density of nodes in the microcellular base station / Picocellcell, femto cell (also known as cell base station, end-to-end node / D2D) and other .5G systems will be more than 10 times higher than the present one, although the ultra-dense network reduces the distance between the end user and the node base station. The spectral efficiency of the network is greatly improved and the system capacity is expanded. However, the number of low-power nodes, the reduction of the distance between nodes, and the increasingly dense deployment of network nodes make the network topology more dense. Because of isomerization and complication, the interference environment becomes more complicated. Therefore, it is important to study the modeling of 5G heterogeneous network interference. The traditional hexagonal mesh model is fixed but not accurate. For the future multilayer isomerization of 5G super-dense networks, a more common method in recent years is to model the network deployment by using Poisson point process in random geometry. It is assumed that the base station positions of each layer conform to the same point process or use a completely random Poisson point process. Due to the edge area of the macro cell (blind area, hot spot area of the macro cell) and D2D communication device to Device, the location of the base station in each layer is assumed to be the same. In special areas such as communication, The distribution of base stations is not the same, and the distribution of user terminals is different in different scenarios. Therefore, the existing simple Poisson point process model will no longer be applicable. In this paper, we use different point process models to simulate different network models for different scenes, and analyze their interference. A two-layer MPP-PHP(Matern Hard-core Point Processes-Poisson Hole process interference model with interlayer correlation (spatial correlation between different types of base stations) is established for the edge region of macro cell. The interference model is based on the traditional hexagonal model and the existing simple MPP-PPP(Poisson Point process model. To carry on the simulation contrast, The accuracy and applicability of the model are verified, and the influence of disturbance distribution and PHP repulsion radius on the system performance is analyzed by simulation. The optimal rejection radius is obtained to make the system performance the best under the given simulation parameters. A two-layer MPP-MCP(Matern cluster process interference model with interlayer correlation (spatial correlation between the same type of base stations) is established for hot spots in macro cell. The model is compared with the traditional hexagonal model and the existing simple PPP-MCP model. The accuracy and applicability of the model are verified, and the interference distribution of the model is simulated and analyzed. Finally, a three-layer MPP-PHP-PPP jamming model with interlayer correlation is established for D2D scene. The accuracy of the model is verified, and the influence of the interference distribution of the model and the distance between D2D pairs on the performance of the system is analyzed. It is concluded that the distance between the D2D pairs has a certain effect on the system performance under the given simulation parameters in this paper. With the increase of distance, the performance of the system will become worse, but the effect is not a big conclusion.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.5

【参考文献】