LTE-Advanced上行MIMO预编码技术研究与实现

发布时间：2018-05-03 09:58

本文选题：LTE-Advanced + 上行多天线技术　；参考：《电子科技大学》2014年硕士论文

【摘要】：目前国内3大运营商都在积极部署LTE商用网络,并陆续开始正式运营。3GPP提出了LTE的演进方案LTE-A,在LTE-A上行引入闭环空分复用传输模式,将原来的上行单天线传输扩展为最高可支持4层的MIMO传输。由于上行传输引入了MIMO技术,需要分析预编码算法对系统性能的影响。本文首先介绍了LTE-A上行物理层基本概念和上行多天线相关技术,为后续章节上行MIMO预编码仿真链路的搭建提供理论支撑。本文研究了LTE-A上行MIMO码本预编码技术,由于LTE-A已经设计好了上行传输预编码码本,目前主要的研究内容为码本选择算法。通常的码本选择算法的处理颗粒是子载波或者带宽子带,反馈数据量比较大。本文分析比较了常见的基于性能度量和基于右奇异矩阵匹配这两类算法在窄带预编码矩阵索引(PMI)和宽带PMI选择上的性能。此外,本文设计了一种基于迫零接收机的码本选择改进算法,通过最小化最大噪声信号功率,该算法可以获得最佳的接收信噪比,在窄带PMI反馈模式下BER与理想的最优预编码器只有约1dB的性能差距。本文还深入研究了LTE-A上行MIMO非码本预编码技术,介绍了基于信道对角化、三角化和栅格缩减辅助这三类预编码信号处理方案,并通过PUSCH基带信号处理平台进行算法仿真。对角化预编码可以消除数据流之间的干扰,但需要使用额外的功率分配算法来平衡“病态”子信道。基于几何均值分解(GMD)或(均匀信道分解)UCD的三角化预编码方案可以获得等增益的子信道,栅格缩减辅助预编码算法利用信道变换提高了传输信道的正交性,因而这两类算法不需要进行额外的功率分配即可获得更好的BER性能。此外,在特征值分解(EVD)预编码方案的基础上,本文改进了基于LDLH分解的预编码方案,该方案基于MMSE准则进行设计,在“注水”功率分配时频谱效率约有0.6bps/Hz的增益。最后,本文基于Xilinx KC705开发板对LTE-A上行预编码算法进行了FPGA设计与实现,并通过功能仿真和搭建的软硬件验证平台验证了算法电路的正确性。时序分析结果表明本文所设计的预编码电路最大工作频率能够达到151.607MHz,满足LTE-A上行预编码的需要。
[Abstract]:At present, the three major domestic operators are actively deploying the LTE commercial network, and begin to formally operate. 3GPP has put forward the evolution scheme of LTE LTE-A. the closed-loop space division multiplexing transmission mode is introduced into the LTE-A uplink. The original uplink single antenna transmission is extended to up to 4 layers of MIMO transmission. Due to the introduction of MIMO technology in uplink transmission, it is necessary to analyze the influence of precoding algorithm on system performance. This paper first introduces the basic concept of LTE-A uplink physical layer and uplink multi-antenna correlation technology, which provides theoretical support for the subsequent chapters of uplink MIMO pre-coding simulation link construction. This paper studies the LTE-A uplink MIMO codebook precoding technology. Because LTE-A has already designed the uplink transmission precoding codebook, the main research content is codebook selection algorithm. The usual codebook selection algorithm deals with sub-carriers or bandwidth subbands, and the feedback data is large. This paper analyzes and compares the performance of two kinds of algorithms based on performance metric and right singular matrix matching in narrow band precoding matrix index (PMI) and wideband PMI selection. In addition, an improved codebook selection algorithm based on zero-forcing receiver is designed in this paper. By minimizing the maximum noise signal power, this algorithm can obtain the best received SNR. In the narrowband PMI feedback mode, the performance gap between BER and ideal optimal precoder is only about 1dB. This paper also studies the LTE-A uplink MIMO non-codebook precoding technology, introduces three kinds of precoding signal processing schemes based on channel diagonalization, triangulation and grid reduction, and simulates the algorithm through the PUSCH baseband signal processing platform. Diagonal precoding can eliminate interference between data streams, but additional power allocation algorithms are needed to balance "ill-conditioned" subchannels. The triangular precoding scheme based on geometric mean decomposition (GMD) or (uniform channel decomposition / UCD) can obtain equal gain subchannels. The raster reduction auxiliary precoding algorithm improves the orthogonality of transmission channels by channel transformation. Therefore, these two algorithms do not need additional power allocation to achieve better BER performance. In addition, on the basis of eigenvalue decomposition (EVD) precoding scheme, this paper improves the precoding scheme based on LDLH decomposition. The scheme is designed based on MMSE criterion, and the spectrum efficiency has the gain of 0.6bps/Hz when "water injection" power allocation. Finally, the FPGA design and implementation of LTE-A uplink precoding algorithm based on Xilinx KC705 development board are carried out, and the correctness of the algorithm circuit is verified by functional simulation and software and hardware verification platform. The timing analysis results show that the maximum working frequency of the designed precoding circuit can reach 151.607 MHz, which meets the need of LTE-A uplink precoding.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.5

【相似文献】