基于高斯树近似的大规模MIMO检测算法

Massive MIMO detection algorithm based on Gaussian tree approximation

大规模MIMO系统中,基站端复杂的接收信号将导致信号检测困难等问题。该文提出一种基于截断诺依曼级数近似求逆高斯树的低复杂度消息传递算法,该算法在高斯树近似消息传递的过程中利用诺依曼级数近似求解协方差矩阵和最小均方误差估计以降低计算复杂度。仿真结果表明,该文算法在检测性能满足系统需求的同时,也有较低的计算复杂度。

In Massive MIMO systems, the complex reception signals at the base station end will lead to the difficulty of signal detection. This paper presents a truncated Neumann series approximation algorithm based on transitive low complexity message inverse Gaussian tree, the algorithm of approximate estimation for covariance matrix and minimum mean square error to reduce the computational complexity by using Neumann series process of message passing in Gaussian tree. The simulation results show that the proposed algorithm has a lower computational complexity while the detection performance meets the requirements of the system.