面向IMT-A的D2D系统通信机制优化的研究

发布时间：2018-05-08 22:20

本文选题：终端直通技术 + 设备发现　；参考：《北京交通大学》2014年硕士论文

【摘要】：未来,IMT-Advanced将提供高达100MHz的带宽来实现高速率传输。为了满足IMT-A的要求,许多新的无线接入技术提交给国际电信联盟。其中,由于能够有效提高频谱效率,Device-to-Device作为一项极有前景的技术已经引起人们越来越多的关注。蜂窝模式下的D2D通信可以有效利用通信双方距离较近的优势,增加资源的利用率,改善蜂窝覆盖范围。相对于传统的蜂窝模式,D2D通信需要新的邻居发现方法、模式选择算法和无线资源管理算法来进一步发掘D2D通信的潜在优势。首先,详细的介绍了蜂窝网络下的D2D通信的各项关键技术,对邻居发现的定义、部署场景以及性能指标逐一做了说明。对现有的具有代表性的几类邻居发现算法进行了详细的分析、对比和讨论。对模式选择中几种常见的通信模式进行详细介绍,并分析了蜂窝网络中引入D2D通信后的带来的干扰,主要是D2D以复用方式共享蜂窝资源带来的复杂干扰。其次,介绍了发现场景的新特点及对发现算法的要求,并研究分析了当前的发现算法。针对当前算法中的不足,本文提出了基于OFDM物理层的邻居发现算法,该算法可以同时应用于有基站覆盖时的场景和无基站覆盖时的场景。而在这两种场景下采用相同物理层信道结构,这样做可以使得大大降低终端实现的复杂度。同时,算法采用了灵活的周期设置,使得每个周期的可用资源可以随着系统中的UE数量而变化。最后,为现有的模式选择场景加入了一种全新的场景：D2D中继场景,并详细描述了D2D中继场景的特点和中继选取的准则,最终针对三种不同的场景给出系统总的架构,构造了以系统吞吐量为目标函数的最优化问题,并给出了模式选择和资源分配算法。首先,在不考虑资源复用的情况下,模式选择与资源分配问题被建模为一个混合整数规划问题,通过松弛信道分配参数,问题被转化为一个凸优化问题,二元分解法被用来求解此凸优化问题。其次,为了充分利用空间复用增益,提出了UE配对资源复用算法,通过复用彼此的资源,使得各自的速率得到提升。仿真的结果表明,我们的算法可以有效提高系统的吞吐量。
[Abstract]:In the future, IMT-Advanced will provide a bandwidth of up to 100MHz to achieve high rate transmission. In order to meet the requirements of IMT-A, many new wireless access technologies are submitted to the International Telecommunication Union. As a result of the ability to effectively improve the spectrum efficiency, Device-to-Device has attracted more and more attention as a promising technology. The D2D communication in the mode can effectively utilize the advantages of the distance between the two parties, increase the utilization of resources and improve the coverage of the cellular coverage. Compared with the traditional cellular mode, D2D communication needs new neighbor discovery methods, pattern selection algorithm and wireless resource management algorithm to explore the potential advantages of D2D communication.
Firstly, the key technologies of D2D communication in cellular network are introduced in detail. The definition of neighbor discovery, deployment scene and performance index are explained one by one. The existing representative neighborhood discovery algorithms are analyzed, compared and discussed in detail. This paper introduces and analyzes the interference caused by the introduction of D2D communication in the cellular network, mainly because of the complex interference caused by D2D sharing resources in the way of multiplexing.
Secondly, the new features of the discovery scene and the requirements for the discovery algorithm are introduced, and the current algorithm is studied and analyzed. In this paper, a neighbor discovery algorithm based on the OFDM physical layer is proposed in this paper. The algorithm can be applied to the scene with the base station coverage and the scene without the base station coverage at the same time. In these two kinds of algorithms, the algorithm can be used in these two kinds of algorithms. The same physical layer channel structure is used in the scene, which can greatly reduce the complexity of the terminal implementation. At the same time, the algorithm uses a flexible cycle setting, making the available resources of each cycle can vary with the number of UE in the system.
Finally, a new scene is added to the existing pattern selection scene: D2D relay scene, and the characteristics of the D2D relay scene and the criteria of relay selection are described in detail. Finally, the overall architecture of the system is given for three different scenarios. The optimization problem of the system throughput as the target function is constructed, and the mode selection and the model selection are given. First, in the case of resource reuse, the problem of pattern selection and resource allocation is modeled as a mixed integer programming problem. By relaxing the channel allocation parameters, the problem is converted to a convex optimization problem. The two element decomposition method is used to solve the convex optimization problem. Secondly, in order to make full use of space reuse to increase the problem. The UE pair resource reuse algorithm is proposed. By reusing each other's resources, each rate is improved. The simulation results show that our algorithm can effectively improve the throughput of the system.

【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.5

【共引文献】