论文部分内容阅读
数字模拟混合预编码可以用较少的射频逼近全数字预编码的性能,可以用来解决毫米波大规模MIMO系统中由于射频链路过多造成的硬件损耗和校准问题。为解决传统混合预编码结构难以实现的缺点,该文的混合预编码研究基于一种简单的固定子连接结构。推导了系统可达速率最大,模拟预编码矩阵应满足的条件,从而将混合预编码矩阵设计问题转化为优化问题。采用鸟群算法(BSA)解决此优化问题,求得最优的预编码矩阵。针对模拟移相器分辨率有限的情况,提出一种直接量化的解决方案和一种基于改进的离散BSA的解决方案。仿真结果表明,所提算法能够基于简单结构实现较好的性能;移相器分辨率有限情况下,所提的两种解决方案都是有效的,且基于离散BSA的方案在分辨率较低时性能更优。
Digital analog hybrid precoding allows the performance of full digital precoding to be approximated with less RF and can be used to solve hardware loss and calibration problems in millimeter-wave large-scale MIMO systems due to excessive RF links. In order to solve the shortcomings of the traditional hybrid precoding structure, the hybrid precoding research based on a simple fixed sub-structure. The conditions that the maximum system reachable rate and the simulated precoding matrix should be satisfied are deduced, so that the design problem of hybrid precoding matrix is transformed into the optimization problem. The bird population algorithm (BSA) is used to solve this optimization problem, and the optimal precoding matrix is obtained. Aiming at the limited resolution of analog phase shifter, a direct quantization solution and a solution based on improved discrete BSA are proposed. The simulation results show that the proposed algorithm can achieve better performance based on the simple structure. Under the limited resolution of the phase shifter, the proposed two solutions are effective, and the scheme based on discrete BSA is in low resolution Better performance.