一种基于施密特正交化的MIMO预编码方法

MIMO precoding method based on Schmidt orthogonalization

提出了一种基于施密特正交化的MIMO(multiple-input multiple-output)预编码矩阵的构造方法。该方法以格拉斯曼空间装箱原理生成的码本为基础,依据信道状态信息(channel state information,CSI),在接收端以最小误比特率(bit error rate,BER)为性能指标,选择最佳匹配当前CSI的预编码矢量,进而将该矢量在格拉斯曼码本中的索引反馈给发送端,基站端根据层映射数目,从格拉斯曼码本中寻找与该矢量最相近的矢量,采用施密特正交化的方法构造出当前预编码矩阵。数值仿真结果表明,依据该方法构造的预编码矩阵,在同等条件下,与3GPP LTE R8标准中根据Householder变换构造的预编码矩阵相比,其BER性能更佳,且该方法实现简单,灵活性好。

In the paper, a MIMO （multiple-input multiple-output） precoding matrix construction method based on Schimidt orthogonalization is proposed, which uses the initial eodebook generated by Grassmannian space packing principle. Accord- ing to the CSI （channel state information）, the receiver chooses a precoding vector from the Grassmannian initial codebook, which can match the CSI best by the criterion of the minimum bit error rate, then feedback the codebook index of the cho- sen precoding vector to the transmitter through the feedback link. Then the base station finds out a most similar vector with the feedback one in the Grassmannian eodebook according to the number of the layer mapping. At last, we construct the re- sultant precoding matrix by Schmidt orthogonalization. Simulation resuhs show that the proposed method has better BER performance than that based on the Householder transform in 3GPP LTE R8, and it is also easy to be realized with better flexibility.