基于细化高斯-赛德尔迭代的线性预编码  

Linear pre-coding based on refined Gauss-Seidel method

作　　者：赵季红 张雯[1] 曲桦 王伟华[2] 张佩杰[1] ZHAO Jihong;ZHANG Wen;QU Hua;WANG Weihua;ZHANG Peijie(School of Communication and Information Engineering, Xi'an University of Posts and Telecommunications, Xi'an 710121, China;School of Electronic and Information Engineering, Xi'an Jiaotong University, Xi'an 710054, China)

机构地区：[1]西安邮电大学通信与信息工程学院,陕西西安710121 [2]西安交通大学电子与信息工程学院,陕西西安710054

出　　处：《西安邮电大学学报》2019年第2期1-5,共5页Journal of Xi’an University of Posts and Telecommunications

基　　金：国家自然科学基金资助项目(61531013);国家重大专项资助项目(2018ZX03001016)

摘　　要：在大规模多输入多输出预编码系统中,为提高经典迭代低复杂度线性预编码方法的收敛率,给出一种基于细化高斯-赛德尔迭代的快速收敛预编码方法。不同于经典迭代只求逆矩阵的对角矩阵,所给方法先求逆矩阵的带状矩阵,再通过迭代实现快速编码。带状矩阵含带更多信息,可降低编码系统的误码率。仿真结果显示,所给方法能有效提高预编码的收敛速率、减少迭代次数并降低误码率。In massive multiple input multiple output systems, to improve the convergence rate of linear pre-coding algorithm based on classical iterative method which is of low complexity, a novel fast convergence linear pre-coding algorithm based on refined Gauss-Seidel method is proposed. Different from the classical algorithm which only uses the diagonal matrix of the inverse matrix, the proposed one firstly finds the banded matrix of the inverse matrix, and then realizes the fast coding by iteration. The banded matrix contains more information of the inverse matrix, which can reduce the bit error rate of the coding system. Simulation results show that the proposed algorithm can achieve lower bit error rate and lower iteration number, but get faster convergence rate of pre-coding.

关 键 词：大规模多输入多输出 线性预编码 迭代法 逆矩阵 高斯-赛德尔迭代 

分 类 号：TN929.53[电子电信—通信与信息系统]