大规模MIMO系统中基于权重二对角迭代的低复杂度预编码

A low complexity precoding based on weighted two diagonal iteration in massive MIMO system

大规模MIMO系统中由于系统下行链路的迫零(zero forcing,ZF)预编码中存在大矩阵求逆运算,随着用户数与天线数的增加,其复杂度随之增加。为了降低复杂度,提出了一种基于雅克比(Jacobi)迭代算法的改进预编码算法,用下二对角矩阵作为迭代矩阵,并且将迭代结果与上一步迭代结果进行权重相加来加速迭代。根据大规模MIMO系统信道矩阵的对角占优特性,将矩阵求逆的诺依曼近似的第一项作为迭代的初始值进一步加速迭代。相比于传统迫零预编码方案,提出的方案可以降低一个量级的算法复杂度,并且保证了预编码方案的性能。

In massive MIMO systems, with the increase of the number of users and antennas, the complexity of the downlink zero forcing（ZF） precoding increases due to the existence of large matrix inversion. In order to reduce the complexity, an improved weighted two diagonal iteration precoding algorithm based on the Jacobi iterative algorithm was proposed. The low-two diagonal matrix was taken as iterative matrix, and the iterative results were weighed addition with the last results to accelerate the convergence speed of the iteration. According to the diagonally dominant characteristic of channel matrix in massive MIMO systems, the first term of the Neumann approximation of matrix inversion was used as the initial value of the iteration to accelerate the iteration further. Compared with the traditional ZF precoding, the proposed scheme can reduce the complexity with an order of magnitude and ensure the performance of the precoding scheme.