视距相关信道下考虑非理想CSI的MIMO收发联合设计

MIMO joint transceiver design in Rice correlated channels with imperfect CSI

MIMO系统收发联合设计的性能依赖于信道状态信息(CSI)的质量。在较高通信频段(如6～8GHz)下,信道的视距分量相对于较低频段显著增加,采用传统的CSI模型无法反映实际的信道特点,从而影响联合设计算法的性能。本文在相关莱斯信道下,提出一种考虑高频段环境下非理想CSI的MIMO系统下行链路收发联合设计算法。论文推导出一种基于训练序列的信道估计误差模型代替传统误差模型,并在此基础上以最小化MSE为目标迭代求解最优预编码/解码矩阵。仿真结果表明,该算法在不同的子流数、视距分量、信道相关性以及信道估计误差环境下,具有较好的误码率性能。

The performance of joint transceiver design in the MIMO systems highly depends on the quality of channel state information （CSI）. Unfortunately, the traditional CSI model can not precisely represent the actual channel characteristics with a relatively high frequency band （eg. 6-8GHz） due to the much higher LOS （Line-of-sight） component .of the channel, which may severely affect the performance of joint-design algorithm. With the consideration of correlated Rice channel, a joint transceiver design is provided in this paper for the downlink MIMO based on imperfect CSI with high frequency band. A model of channel-estimation errors based on training sequence is deduced to replace the traditional one. The precoding ＆ decoding matrices are obtained iteratively according to the minimum MSE criterion. The analysis results indicate that the algorithm have a good BER performance with different substream numbers, LOS components, the degree of channel correlation and channel estimation errors.