大规模MIMO系统下行预编码传输方法研究

发布时间：2018-03-15 05:21

本文选题：大规模MIMO　切入点：下行预编码　出处：《东南大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着人们对无线通信业务需求的增长,大规模多输入多输出(Multiple Input Multiple Output,MIMO)技术得到了学术界和工业界越来越广泛的关注。大规模MIMO技术中基站侧配备数量巨大的天线,并且在同一时频资源上服务多个用户,可以有效地提高系统频谱效率以及数据传输的可靠性。在非协作多用户MIMO系统中,预编码是下行链路数据传输的关键技术之一,本文针对大规模MIMO下行预编码方法展开研究。首先,对传统的MIMO预编码方法进行了研究,并提出了一种基于信道格拉姆矩阵特征值分解(Eigenvalue Decomposition,EVD)的最小均方误差(Minimum Mean Square Error,MMSE)预编码方法。传统MMSE预编码的性能受限于多用户信道之间的相关性,用户信道之间的强相关性会严重影响预编码方法的性能。对每一个用户信道进行接收端解相关预处理,极大地降低多用户信道之间的相关性,因而提出的方法能够有效地提高预编码的性能。仿真结果表明,基于信道格拉姆矩阵特征值分解的MMSE预编码较传统预编码具有更优的和速率性能。接着,针对大规模MIMO两级预编码,提出一种用户调度与波束分配算法。在大规模MIMO频分双工系统中,由于信道估计与信道反馈的开销与基站端天线数目成正比,在信道相关时间有限的情况下,准确地估计出完全的瞬时信道信息是非常困难的。利用波束域信道的稀疏特性,提出一种两级预编方法:第一级预编码根据统计信道信息进行用户的调度与波束的分配;第二级预编码对瞬时等效信道信息进行正则化迫零(Regularized Zero Forcing,RZF)预编码,从而进一步抑制用户间数据干扰。仿真结果表明,基于最大化两级预编码和速率准则,提出的用户调度与波束分配算法,不仅能够极大地降低等效信道维度,减少信道估计与反馈的开销,而且能在不损失系统性能的情况下保持合理的算法复杂度。最后,针对毫米波大规模MIMO系统,提出了一种基站侧与用户侧联合用户调度与射频通道分配算法。在毫米波大规模MIMO系统中,高频段电磁波的短波长特性使得用户端设备可以配备大量的天线。此外,基站与接收端之间的射频通道数目非常少,因而射频通道在基站侧和用户侧均呈现出高稀疏特性。利用射频通道的稀疏特性,进行基站侧和用户侧的联合射频通道调度,既可以极大地压缩信道维度,减少信道估计的开销,同时也可以增大功率效率,提高系统通信性能。算法首先进行基站侧的射频通道选择,以最大化和速率为准则,将和速率最大化问题转化为等效的凸差函数规划问题,利用凸优化理论求解得到基站侧射频通道;在得到基站侧射频通道的基础上,再次以最大化和速率为准则,利用贪婪算法选择出用户侧的射频通道。经过两个步骤的处理,可以分别得到基站侧和用户侧的射频通道。仿真结果表明,提出的算法较一般预编码方法具有更优的和速率性能,且极大地降低了信道维度,减小了信道估计的开销。
[Abstract]:With the increasing demand for wireless communication business, large-scale multi input multi output (Multiple Input Multiple Output MIMO) technology has been in academia and industry more and more attention. Mass base in MIMO technology with a huge number of antenna side, and a number of service users in the same time-frequency resources, can effectively to improve the system spectrum efficiency and reliability of data transmission. In the non cooperative multi-user MIMO system, pre encoding is one of the key technologies of the downlink data transmission, based on large-scale MIMO downlink pre encoding method to develop research. Firstly, MIMO pre encoding of the traditional methods are studied, and propose a channel characteristic value decomposition based on the Gram matrix (Eigenvalue Decomposition, EVD) of the minimum mean square error (Minimum Mean Square Error, MMSE) pre encoding method. The performance of traditional MMSE precoding by Limited to the correlation between the multiuser channel, strong correlation between user channel will seriously affect the performance of the pre encoding method. Receiver solution pretreatment for each user channel, greatly reduce the correlation between multiple users of the channel, so the proposed method can effectively improve the performance of the pre encoding. The simulation results show that the the channel characteristics of Gram matrix value decomposition MMSE precoding precoding is better than the traditional and rate based on performance. Then, for large scale MIMO two pre encoding, proposes a user scheduling and beam allocation algorithm. MIMO frequency division duplex system in large scale, because the channel estimation is proportional to the overhead and the base station antenna number channel feedback, in the channel with limited time, accurate estimation of the instantaneous channel information completely is very difficult. The sparse characteristics of the beam domain channel Two, put forward a method of pre: the distribution of the first pre encoding scheduling and user according to the beam statistical channel information; second pre encoding of instantaneous equivalent channel information is regularized zero forcing (Regularized Zero, Forcing, RZF) - encoding, data to further suppress the interference between users. The simulation results show that the maximum two pre encoding and rate based on the criterion of user scheduling and beam allocation algorithm is proposed, which can not only reduce the dimension of the equivalent channel, reduce the channel estimation and feedback overhead, and without loss of energy in the system can keep the reasonable complexity. Finally, according to the millimeter wave of large-scale MIMO system, is put forward a base station side and user side joint user scheduling and radio channel allocation algorithm. In the millimeter wave in the large scale MIMO system, short wavelength characteristics of high-frequency electromagnetic wave so that the terminal user equipment With the large number of antennas. In addition, the number of radio channel between the base station and the receiving end is very small, so the radio channel at the base station side and user side showed high sparsity. Using the sparsity of the RF channel, combined with radio frequency channel scheduling base station side and user side, which can greatly reduce the dimension compression channel. The channel estimation overhead, but also can increase the power efficiency, improve the communication performance of the system. The RF channel algorithm first base station selection, and to maximize the rate as the criterion, and the rate maximization problem into the equivalent convex function programming problem, optimization theory is obtained by convex side of base station RF channel; in based on base station RF channel, again to maximize the sum rate criterion, using the greedy algorithm to select a radio channel user side. After processing two steps, respectively. The radio channel of base station and user side is obtained. Simulation results show that the proposed algorithm has better and rate performance than the general precoding method, and greatly reduces the channel dimension and reduces the cost of channel estimation.

【学位授予单位】：东南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN919.3

【相似文献】