基于重叠子阵的毫米波大规模MIMO混合预编码算法

Overlapped Subarray Based Hybrid Precoding for mmWave Massive MIMO

基于全连接的毫米波大规模MIMO系统下行链路通信具有高的硬件成本及能量消耗,为解决此问题,同时为解决MIMO系统基站和移动端均混合设计且多流传输的优化问题,提出一种基于重叠子阵列架构(OSA)的混合预编码算法,利用通用子阵列架构在发射机处执行模拟预编码和数字预编码。模拟预编码算法采用改进的矩阵广义低秩近似来获取RF预编码矩阵/组合矩阵,以获取大规模MIMO系统所提供的高阵列增益。在得到模拟预编码矩阵后,进而得到系统的射频等效信道,然后采用块对角化(Block Diagonalization,BD)算法求取数字预编码矩阵。仿真结果表明,针对OSA设计提出的改进的矩阵广义低秩近似混合预编码算法,减少迭代计算次数,与现有的基于OAS架构的混合预编码算法相比,具有更高的频谱效率和较低的复杂度,更接近全数字预编码算法性能。

Downlink communication in millimeter wave massive MIMO system requires significantly complex hardware and high energy consumption.To address this issue,along with the optimization problem of mixed base station and mobile terminal design and multi-stream transmission in MIMO systems,we propose a hybrid precoding scheme based on the overlapped subarray architecture(OSA),where analog precoding and digital precoding are implemented at the transmitter using the sub-array architecture.Specifically,analog precoding algorithm uses modified matrix generalized low rank approximation to obtain RF precoding matrix or combination matrix,and thus obtain high array gain provided by large-scale MIMO system.AS a result,the equivalent radio frequency channel of the system can be calculated.Then the block diagonalization(BD)algorithm is used to obtain the digital precoding matrix.Simulation results show that the modified matrix generalized low rank approximate hybrid precoding algorithm for OSA design can reduce the number of iterations.Compared with the existing hybrid precoding algorithm,it achieves higher spectral efficiency and lower complexity.