基于空间相关性的大规模多用户MIMO预编码

Research on massive multi-user MIMO precoding based on spatial correlation

由于FDD(frequency-division duplex)系统的信道不具有互易性,使得大规模MIMO(multiple-input multiple-output)系统发射端对多天线信道信息的获取变得十分困难。为了解决基站获取信道状态信息的问题,可以采用双重结构预编码方案,将对长期CSIT处理的波束成形和随后对短期CSIT处理的线性预编码进行级联,可以有效地降低反馈开销。预编码矩阵由基于空间相关性的预波束成形矩阵和基于短期CSIT的传统预编码矩阵决定。通过将具有空间相关性的用户进行分组,并将分组后的用户的信道矩阵与基于空间相关性的相同的预处理矩阵相乘,使预编码信号空间的维度降低。

Because the channel of FDD system lacks reciprocity, it becomes very difficult to obtain the multi-antenna channel information by the transmitting terminal of the massive MIMO system. In order to solve the problem that the base station obtains the state information of channel, a dual-structure precoding scheme can be used. The longterm CSIT beamforming and the linear preeoding of the short-term CSIT processing are cascaded to effectively reduce the feedback overhead. The precoding matrix is determined by the pre-beamforming matrix based on spatial correlation and the traditional precoding matrix based on short-term CSIT. By grouping the users with spatial correlation, multiplying the channel matrix of the grouped users with the same pre-processing matrix based on spatial correlation, the dimensionality of the precoding signal space can be reduced.