3D MIMO系统中一种有限反馈开销降低方法

针对三维多输入多输出（3D MIMO）预编码系统的高反馈开销问题，结合信道预测，提出了一种基于时频域二维分簇的有限反馈开销降低方法。该方法利用信道的时频域相关特性，用户端根据与当前帧中数据块所对应的信道状态信息（CSI），采用自回归（AR）模型来预测与后续帧中数据块相对应的3D CSI，进而对预测得到的CSI进行时域与频域二维分簇处理，最后将与每个反馈数据块相对应的预编码矩阵索引反馈给基站。仿真结果表明：该方法在保证系统性能的同时，不但大幅度地降低3D MIMO多用户系统的反馈开销，而且有效地改善了反馈延迟对系统性能所造成的影响。

Overhead Reduction Method with Limited Feedback for 3D MIMO Systems

Combined the channel perdition, we propose a new limited feedback method based on the two-dimensional clustering to reduce the high feedback overhead for three dimension Multiple Input Multiple Output system. The algorithm utilizes channel correlation in both time domain and frequency domain. According to the current 3D channel state information, the users adopt the autoregressive model to predict the CSI associated with the subsequent frame, and the predicted CSI is clustered in the time and frequency domain. Finally, the indexes of precoding matrix associated each data block are fed back to Base Station. Simulation results show that the feedback overhead is greatly reduced and the performance degradation caused by large feedback delay can be well mitigated.