A new MIMO precoding method is proposed based on maximizing the minimum Euclidean distance between points in the received signal space for triple-stream transmission. Under a real-valued rotation matrix constraint, the proposed algorithm solves the non-convex optimization problem with two steps from the geometric viewpoint, named offline step and online step, respectively. In the offline step, all feasible surfaces are collected by expanding quadric surfaces within 3-dimensional space. In the online step, the precoder matrix is constructed with respect to the unique feasible surface chosen by particular channel implementation. Meanwhile, a closed-form expression of the minimum distance is also provided in this paper. Simulation results validate that the proposed method not only outperforms existing schemes with low computational complexity, but also has superior performance of more than 3 dB in terms of bit-error-rate (BER) with increase of the quadrature amplitude modulation (QAM) constellations.%基于最大化接收信号的最小欧氏距离准则,提出一种适用于传输三路数据流的多输入多输出预编码新算法.该算法利用实参数旋转矩阵约束,从几何角度将非凸预编码优化问题的求解转化为离线计算和在线计算两步实施,并获得表征最小欧氏距离的闭合表达式,其中离线计算部分通过搜索三维空间中依特定条件膨胀的二次曲面并获得所有可行曲面,而在线计算部分则根据信道条件选择可行曲面,从而构造出相应的预编码矩阵.仿真实验结果和分析表明,该算法相比于其他同类算法不仅具有较低的计算复杂度,而且随着正交幅度调制星座图的增大可以获得超过3 dB的误比特率性能增益.
展开▼
