本文分析了预编码算法现状,提出了SINR功率分配方式,它在整个预编码中计算复杂度相对于注水原理是比较低的。研究了它在一种算法中的应用,该算法通过SVD奇异值分解得到预编码矩阵,用SINR来实现功率分配。由于避免了注水原理的迭代运算,因此功率分配计算复杂度有所降低。与此同时,系统和容量与误码率相对于传统平均功率分配方式都有了明显的提升。 This paper analyzes the status quo of the pre-coding algorithm, and proposes a SINR power allocation mode, which has lower computational complexity than the flooding principle in the entire precoding. It is studied in an algorithm that uses SVD singular value decomposition to obtain a precoding matrix and uses SINR to achieve power allocation. Due to the avoidance of the iterative computation of the waterflooding principle, the computational complexity of the power allocation is reduced. At the same time, the system and capacity and bit error rate have been significantly improved over the traditional average power allocation.