用于大规模MIMO系统下行链路的低复杂度预编码算法  

Low-complexity precoding in downlink of massive MIMO systems

作　　者：朱益胜 姚建国[1,2] 

机构地区：[1]南京邮电大学通信与信息工程学院,南京210003 [2]南京邮电大学江苏省无线通信重点实验室,南京210003

出　　处：《计算机应用》2017年第A02期20-23,共4页journal of Computer Applications

摘　　要：在传统的大规模多输入多输出(MIMO)系统中,预编码在确定大规模MIMO系统下行链路可实现速率方面起重要作用;然而大多只考虑了单小区系统,忽略了小区间干扰。针对大规模MIMO系统在多小区中的问题,提出了一种优化的连续迫零脏纸预编码算法。首先,讨论了一种基于脏纸编码和迫零的低复杂度方法,它结合了简化的QR分解和正交投影。其次,综合考虑了小区内和小区间干扰的下行链路的和速率优化问题;最后,使用凸共轭将问题转化为无约束双重问题,应用具有低复杂度的拟牛顿算法来找到最优迭代解。仿真实验中,在天线数相同时,除了脏纸编码算法外,提出的算法比其他算法拥有更高的吞吐量,而且在性能上比其他传统算法提升了约10%,并且具有更快的收敛性。仿真分析表明,该预编码算法优于其他常规预编码方法。In traditional massive Multiple Input Multiple Output（ MIMO） systems, precoding plays an important role in determining the achievable rate of downlink in massive MIMO systems. However, only the single-cell systems are considered,and inter-cell interference is ignored. To solve the problem of massive MIMO system in multi-cell, an optimized successive zero forcing-dirty paper coding algorithm was proposed. First, a low-complexity method was discussed based on dirty paper coding and zero-forcing, which combines simplified QR decomposition and orthogonal projection. Secondly, the downlink sum rate optimization problems of intra-cell and inter-cell interference were considered. Finally, the convex conjugate was used to convert the problem into an unconstrained dual problem, and the quasi-Newton algorithm with low complexity was used to find the optimal iterative solution. In the simulation experiment, besides the dirty paper coding algorithm, the proposed algorithm has higher throughput than other algorithms when the antennas are equal, but the performance was increased by about 10%compared to others, and has a faster convergence. Simulation analysis shows that the proposed precoding algorithm is superior to other conventional precoding methods.

关 键 词：低复杂度 大规模多输入多输出 和速率 预编码 拟牛顿算法 

分 类 号：TN929.5[电子电信—通信与信息系统]