多小区Massive MIMO系统低复杂度ZF线性检测算法  

Low-Complexity ZF Linear Detection Algorithm for Multi-Cell Massive MIMO System

作　　者：张瑞欣 曹海燕[1] 谢时埸 王秀敏[2] 

机构地区：[1]杭州电子科技大学,浙江杭州310018 [2]中国计量大学,浙江杭州310018

出　　处：《通信技术》2017年第10期2250-2254,共5页Communications Technology

基　　金：国家自然科学基金资助项目(No.61501158;No.61379027);浙江省自然科学基金资助项目(No.LY14F010019;No.LQ15F01004)~~

摘　　要：针对多小区Massive MIMO上行链路系统中ZF线性检测涉及到大矩阵求逆而具有高复杂度的问题,提出了一种低复杂度的ZF检测算法。在考虑各小区信道状态信息已知条件下,利用SVD分解的方法求解干扰消除矩阵,从而将多小区信号接收模型等效转化为单小区模型,然后再利用ZF检测算法。但是,在ZF检测算法中涉及到大矩阵求逆运算的复杂度为O(K^3),其中K为本小区中的用户数。为了降低直接求逆的高复杂度,提出了将大矩阵分解为对角矩阵和空心矩阵之和,并采用诺依曼级数近似且通过优化展开项因子,使所提算法在性能损失很少的情况下复杂度降低了一个数量级为O(K^2)。此外,通过仿真实验验证了理论推导与分析的有效性。ZF（Zero Forcing） detection algorithm has high complexity since the large matrix inversion in multi-cell Massive Multiple Input Multiple Output（MIMO） uplink system is involved.In view of this problem,a low-complexity ZF detection algorithm is proposed.Considering the channel state information of each cell,the SVD decomposition method is used to solve the interference cancellation matrix.In this way,the multi-cell signal reception model is transformed into a single-cell model equivalently,and then the ZF detection algorithm could be used.However,the inverse-operation complexity of large matrices involved in the ZF detection algorithm is O（K^3）,where K is the number of users in this cell.In order to reduce the high complexity of direct inversion,the large matrix is decomposed into the sum of diagonal matrix and hollow matrix.By using the Neumann series approximation and by optimizing the expansion term factor,the complexity of the proposed algorithm is reduced by an order of magnitude O（K^2） in the case of little performance-loss.In addition,simulation experiment also indicates the validity of theoretical derivation and analysis.

关 键 词：多小区Massive MIMO ZF检测算法 诺依曼级数近似 低复杂度 

分 类 号：TN92[电子电信—通信与信息系统]