ABSTRACT

Conventional multiple-input multiple-output (MIMO) systems operating at sub-6 GHz frequencies generally employ fully digital precoding which requires dedicated radio frequency (RF) chain for each antenna element. However, this architecture is infeasible for the millimetre wave (mmWave) massive MIMO systems due to excessive power consumption and space constraints posed by large number of RF chains. Hence, hybrid precoding architecture, which utilises small size digital precoder followed by large size analog precoder, is employed for such systems. Due to the constant-magnitude hardware constraint on the phase shifter elements of analog precoder, efficient design of hybrid precoders is a challenge. In this paper, a low-complexity non-iterative hybrid precoder/combiner design is proposed for a single-user mmWave massive MIMO system. The analog RF precoder vectors are selected non-iteratively from the set of transmit array response vectors such that they are ‘aligned’ with the optimal precoder of the channel. Then, for the fixed analog precoder, the digital baseband precoder is evaluated using least square solution. Matrix operations like singular value decomposition, correlation and inversion are required only once. A similar approach is adopted for the design of hybrid combiners of the receiver. Simulation results show effectiveness of the proposed method in terms of spectral efficiency maximisation.

摘要

在低于 6 GHz 频率下运行的传统多输入多输出 (MIMO) 系统通常采用全数字预编码，这需要为每个天线元件提供专用射频 (RF) 链。然而，由于大量射频链带来的过度功耗和空间限制，这种架构对于毫米波 (mmWave) 大规模 MIMO 系统是不可行的。因此，这种系统采用了混合预编码架构，该架构使用小尺寸数字预编码器，然后是大尺寸模拟预编码器。由于对模拟预编码器的移相器元件的恒定大小的硬件限制，混合预编码器的高效设计是一个挑战。在本文中，针对单用户毫米波大规模 MIMO 系统提出了一种低复杂度的非迭代混合预编码器/组合器设计。模拟RF预编码器矢量从发射阵列响应矢量集合中非迭代地选择，使得它们与信道的最佳预编码器“对齐”。然后，对于固定模拟预编码器，使用最小二乘法评估数字基带预编码器。像奇异值分解、相关和求逆这样的矩阵运算只需要一次。接收机的混合合路器的设计采用了类似的方法。仿真结果表明所提出的方法在频谱效率最大化方面的有效性。像奇异值分解、相关和求逆这样的矩阵运算只需要一次。接收机的混合合路器的设计采用了类似的方法。仿真结果表明所提出的方法在频谱效率最大化方面的有效性。像奇异值分解、相关和求逆这样的矩阵运算只需要一次。接收机的混合合路器的设计采用了类似的方法。仿真结果表明所提出的方法在频谱效率最大化方面的有效性。