基于新型网络架构的5G系统信道估计和预编码技术研究

发布时间：2018-05-29 11:02

本文选题：信道状态信息 + 信道估计　；参考：《电子科技大学》2017年硕士论文

【摘要】：近年来,随着智能手机的进一步普及,用户们对于高速数据业务的需求呈爆炸式增大,这些需求促进了我们对第五代移动通信系统(5G)的研究。目前,中国、欧盟和美国都开始了对5G系统关键技术的研究,并且预计将在2020年开始商用。相较于现有的第四代移动通信系统(4G)系统,5G系统需要在系统容量和频谱利用率上,进一步提高系统的传输性能。故而,5G系统需要应用一些更加高效的技术手段,比如Massive-MIMO技术、信道估计技术、预编码技术等等。Massive-MIMO系统中的复杂的、大量的实时数据计算,就会极大增加系统的数据处理时延,而未来5G系统的一个非常重要的需求就是数据处理中的超低时延。ITU、IMT-2020等国内外5G研究机构均对5G提出了毫秒级的端到端时延要求,即理想情况下端到端时延为1ms,典型端到端时延为5-10ms左右。为了应对未来5G系统对于极低时延的需求,在第二章中,我们介绍了新型软件定义协议(Software-Defined Protocol,SDP)网络架构的基本思想,并在此SDP网络架构下提出了软件定义扇区(Software Defined Sector,SDS)模式,并对这种软件定义扇区模式的计算复杂度性能优势进行仿真验证;最后介绍了基于软件定义扇区模式的时分导频帧格式,并将对其进行了计算复杂度的仿真分析。Massive-MIMO技术有着很高的频谱利用率、很好的鲁棒性和很高的传输速率等优势,故而成为下一代蜂窝移动通信系统(5G)中非常被看好的技术之一。移动信道的信道状态信息的获得是Massive-MIMO系统中的关键问题。在第三章中,我们首先介绍了传统MIMO系统中的四种经典信道估计算法,并分析它们在MassiveMIMO系统中的性能和不足;接着提出了一种新型软件定义扇区模式信道估计算法,即SDS-MMSE信道估计算法,并对其进行了性能仿真和结果分析。本文提出的SDS-MMSE信道估计算法可以获得接近理想MMSE信道估计的MSE性能表现,而不需要额外的开销。我们还得到了本文所提算法的MSE分析表达式,这将帮助我们进行进一步的系统设计和性能估计。在Massive-MIMO系统中,由于渐进正交的信道特性,线性预编码技术可以实现近似通道容量的性能表现。在第四章中,我们首先介绍了Massive-MIMO系统中三种经典预编码算法,然后分析它们在Massive-MIMO系统中的性能、表现和问题。接着,我们提出了一种软件定义扇区模式下的基于对称逐步超松弛(Symmetric Successive Over Relaxation,SSOR)的预编码算法,即SDS-SSOR预编码算法,然后对其进行了性能仿真并分析仿真结果。本文所提的SDS-SSOR预编码算法不仅可以大幅度地降低计算复杂度,并且可以实现接近传统ZF预编码的最优的性能表现。同时,利用Massive-MIMO系统的信道具备渐进正交的信道特性,我们提出了一种简单方法,来计算SDS-SSOR预编码算法中的最佳松弛参数。
[Abstract]:In recent years, with the further popularization of smart phones, users' demand for high-speed data services has increased explosively, which has promoted our research on the fifth generation mobile communication system (5G). At present, China, the European Union and the United States have begun to study the key technologies of the 5G system and are expected to begin commercial use in 2020. Compared with the existing fourth generation mobile communication system (4G), the 5G system needs to further improve the transmission performance in terms of system capacity and spectrum efficiency. Therefore, some more efficient techniques, such as Massive-MIMO, channel estimation, precoding and so on, need to be applied to the 5G system. The complex and large amount of real-time data computation in the Massive-MIMO system will greatly increase the data processing delay of the system. One of the most important requirements of 5G system in the future is that 5G research institutions, such as the ultra-low delay, ITUUUIMT-2020 and so on, have put forward the requirement of 5G end-to-end delay in millisecond. The ideal end-to-end delay is 1 Ms, and the typical end-to-end delay is about 5-10ms. In order to meet the demand of 5G system for very low delay in the future, in chapter 2, we introduce the basic idea of a new software definition protocol (Software-Defined Protocol) network architecture, and propose a software definition sector Defined Defined (SDS) model under this SDP network architecture. Finally, the time-division pilot frame format based on the software definition sector mode is introduced. The Massive-MIMO technology has the advantages of high spectral efficiency, good robustness and high transmission rate, so it has become one of the most promising technologies in the next generation cellular mobile communication system. The acquisition of channel state information in mobile channels is a key problem in Massive-MIMO systems. In the third chapter, we first introduce four classical channel estimation algorithms in traditional MIMO system, and analyze their performance and shortcomings in MassiveMIMO system, then we propose a new software defined sector mode channel estimation algorithm. That is, SDS-MMSE channel estimation algorithm, and its performance simulation and results analysis. The proposed SDS-MMSE channel estimation algorithm can obtain the MSE performance close to the ideal MMSE channel estimation without additional overhead. We also get the MSE analysis expression of the proposed algorithm, which will help us to carry out further system design and performance estimation. In Massive-MIMO systems, linear precoding can achieve the performance of approximate channel capacity due to the asymptotic orthogonal channel characteristics. In chapter 4, we first introduce three classical precoding algorithms in Massive-MIMO system, and then analyze their performance, performance and problems in Massive-MIMO system. Then, we propose a kind of precoding algorithm based on symmetric step by step overrelaxation Successive Over Relaxation (SSOR) in sector mode defined by software, that is, SDS-SSOR precoding algorithm. Then, the performance of the algorithm is simulated and the simulation results are analyzed. The SDS-SSOR precoding algorithm proposed in this paper can not only greatly reduce the computational complexity, but also achieve the best performance close to the traditional ZF precoding. At the same time, we propose a simple method to calculate the optimal relaxation parameters in the SDS-SSOR precoding algorithm because the channel of Massive-MIMO system is asymptotically orthogonal.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.5

【相似文献】