基于块对角化的格基规约GMD矢量预编码

研究了对多用户多输入多输出下行链路进行块对角化后，使用格基规约算法的几何均 值分解矢量预编码的实现方法。根据块对角化思想将多用户多天线信道分解为等价并行子信 道，基于等价子信道给出了单个用户的几何均值分解矢量预编码的传输方案，通过使用格基 规约算法分别结合连续干扰消除和垂直分层空时编码两种方法，求解矢量预编码中的扰动矢 量。仿真表明，提出的方法误码率性能优于块对角化矢量预编码算法2 dB以上，而且能在不 降低系统性能的前提下降低计算复杂度。

Lattice reduction GMD vector precoding based on block diagonalization

Based on the block diagonalization (BD) for multiuser Multiple-input Mu ltiple-output (MIMO) downlink channel, a vector precoding (VP) combined with ge ometric mean decomposition (GMD) based on lattice reduction (LR) algorithm is st udied. The multi-user multi-antenna channel is decomposed into parallel effect ive sub-channels, and then a GMD VP transmission scheme for single user is pres ented based on the above effective sub-channel. The LR algorithm combined with successive interference cancellation (SIC) and Vertical Bell Laboratories Layered Sp ace-Time(VBLAST) is exploited, so that the perturbation vector is obtained. Si mulation results show that the bit error rate (BER) of proposed method outperform s that of block diagonalized vector precoding at least 2dB, and the LR algorithm can reduce the computation complexity with the same system performance.