Efficient Low-Complexity Antenna Selection Algorithms in Multi-User Massive MIMO Systems With Matched Filter Precoding
Efficient Low-Complexity Antenna Selection Algorithms in Multi-User Massive MIMO Systems With Matched Filter Precoding
复制标题
DOI：
10.1109/tvt.2020.2966565
复制
发表时间：
2020-01
影响因子：
6.8
通讯作者：
Zaid Abdullah;C. Tsimenidis;Gaojie Chen;M. Johnston;J. Chambers
中科院分区：
文献类型：
--
作者：
Zaid Abdullah;C. Tsimenidis;Gaojie Chen;M. Johnston;J. Chambers
针对下行多用户（MU）大规模多输入多输出（M-MIMO）系统，提出了两种基于匹配滤波器（MF）预编码的高效低复杂度天线选择（AS）算法。这两种算法在迭代选择过程中避免了向量乘法，以降低复杂度。第一种算法考虑在相同的时频资源上，基站有$N$个天线，为$K$个单天线用户服务，将可用天线分成$K$组，第k$组包含对第k$个用户具有最大信道范数的$N/K$个天线。因此，可以通过仅从第k$个组中选择天线的子集来最大化第k$个用户的信号与干扰加噪声比（SINR），从而导致搜索空间减小了K$倍。第二种算法是半盲干扰抑制方法，其仅依赖于干扰项的符号，并且在每次迭代时，将选择抑制最大数量的干扰项的天线。我们所提出的方法的性能进行评估下完美和不完美的信道状态信息（CSI），并与其他低复杂度的AS计划相比，可实现的和速率以及能源效率。特别地，当信噪比（SNR）为10 dB时，并且对于具有20 MHz带宽的系统，所提出的方法优于其中对于第一和第二所提出的算法分别使用所有天线的情况108.8和49.2Mbps，假定BS具有完善的CSI知识并且配备有256个天线，其中64个被选择来服务8个单天线用户。
In this work, two efficient low complexity Antenna Selection (AS) algorithms are proposed for downlink Multi-User (MU) Massive Multiple-Input Multiple-Output (M-MIMO) systems with Matched Filter (MF) precoding. Both algorithms avoid vector multiplications during the iterative selection procedure to reduce complexity. Considering a system with $N$ antennas at the Base Station (BS) serving $K$ single-antenna users in the same time-frequency resources, the first algorithm divides the available antennas into $K$ groups, with the $k$th group containing the $N/K$ antennas that have the maximum channel norms for the $k$th user. Therefore, the Signal-to-Interference plus Noise Ratio (SINR) for the $k$th user can be maximized by selecting a subset of the antennas from only the $k$th group, thereby resulting in a search space reduction by a factor of $K$. The second algorithm is a semiblind interference rejection method that relies only on the signs of the interference terms, and at each iteration the antenna that rejects the maximum number of interference terms will be selected. The performance of our proposed methods is evaluated under perfect and imperfect Channel State Information (CSI) and compared with other low complexity AS schemes in terms of the achievable sum rate as well as the energy efficiency. In particular, when the Signal-to-Noise Ratio (SNR) is 10 dB, and for a system with 20 MHz of bandwidth, the proposed methods outperform the case where all the antennas are employed by 108.8 and 49.2 Mbps for the first and second proposed algorithms, respectively, given that the BS has perfect CSI knowledge and is equipped with 256 antennas, out of which 64 are selected to serve 8 single-antenna users.